Dissertations / Theses on the topic 'COLLABORATIVE FILTERING ALGORITHMS'
To see the other types of publications on this topic, follow the link: COLLABORATIVE FILTERING ALGORITHMS.
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 48 dissertations / theses for your research on the topic 'COLLABORATIVE FILTERING ALGORITHMS.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
1
Hansjons, Vegeborn Victor, and Hakim Rahmani. "Comparison and Improvement Of Collaborative Filtering Algorithms." Thesis, KTH, Skolan för datavetenskap och kommunikation (CSC), 2017. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-209468.
Full textAbstract:
Recommender Systems is a topic several computer scientists have researched. With today’s e-commerce and Internet access, companies try to maximize their profit by utilizing var- ious recommender algorithms. One methodology used in such systems is Collaborative Filtering. The objective of this paper is to compare four algorithms, all based on Collaborative Filtering, which are k-Nearest-Neighbour, Slope One, Singular Value Decomposition and Average Least Square algorithms, in order to find out which algorithm produce the best pre- diction rates. In addition, the paper will also use two mathematical models, the Arithmetic Median and Weighted Arithmetic Mean, to determine if they can improve the prediction rates. Singular Value Decomposition performed the best out of the four algorithms and Aver- age Least Square performed the worst. However, the Arithmetic Median performed slightly better than Singular Value Decomposition and the Weighted Arithmetic Mean performed the worst.Rekommendationssystem är ett ämne som många datatekniker har forskat inom. Med dagens e-handel och Internetåtkomst, så försöker företag att maximera sina vinster genom att utnyttja diverse rekommendationsalgoritmer. En metodik som används i sådana system är Collaborative Filtering. Syftet med denna uppsats är att jämföra fyra algoritmer, alla baserade på Collaborati- ve Filtering, vilket är k-Nearest-Neighbour, Slope One, Single Value Decomposition och Average Least Square, i syfte att ta reda på vilken algoritm som producerar den bästa be- tygsättningen. Uppsatsen kommer även använda sig av två olika matematiska modeller, Aritmetisk Median och Viktad Aritmetisk Median, för att ta reda på om dom kan förbättra betygsättningen. Single Value Decomposition presterade bäst medan Average Least Square presterade sämst av de fyra algoritmerna. Däremot presterade Aritmetiska Median en aning bättre än Single Value Decomposition och Viktad Aritmetisk Median presterade sämst.
2
Anne, Patricia Anne. "Semantically and Contextually-Enhanced Collaborative Filtering Recommender Algorithms." Thesis, University of Ulster, 2009. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.516289.
Full text
3
Casey, Walker Evan. "Scalable Collaborative Filtering Recommendation Algorithms on Apache Spark." Scholarship @ Claremont, 2014. http://scholarship.claremont.edu/cmc_theses/873.
Full textAbstract:
Collaborative filtering based recommender systems use information about a user's preferences to make personalized predictions about content, such as topics, people, or products, that they might find relevant. As the volume of accessible information and active users on the Internet continues to grow, it becomes increasingly difficult to compute recommendations quickly and accurately over a large dataset. In this study, we will introduce an algorithmic framework built on top of Apache Spark for parallel computation of the neighborhood-based collaborative filtering problem, which allows the algorithm to scale linearly with a growing number of users. We also investigate several different variants of this technique including user and item-based recommendation approaches, correlation and vector-based similarity calculations, and selective down-sampling of user interactions. Finally, we provide an experimental comparison of these techniques on the MovieLens dataset consisting of 10 million movie ratings.
4
Rault, Antoine. "User privacy in collaborative filtering systems." Thesis, Rennes 1, 2016. http://www.theses.fr/2016REN1S019/document.
Full textAbstract:
Les systèmes de recommandation essayent de déduire les intérêts de leurs utilisateurs afin de leurs suggérer des items pertinents. Ces systèmes offrent ainsi aux utilisateurs un service utile car ils filtrent automatiquement les informations non-pertinentes, ce qui évite le problème de surcharge d’information qui est courant de nos jours. C’est pourquoi les systèmes de recommandation sont aujourd’hui populaires, si ce n’est omniprésents dans certains domaines tels que le World Wide Web. Cependant, les intérêts d’un individu sont des données personnelles et privées, comme par exemple son orientation politique ou religieuse. Les systèmes de recommandation recueillent donc des données privées et leur utilisation répandue nécessite des mécanismes de protection de la vie privée. Dans cette thèse, nous étudions la protection de la confidentialité des intérêts des utilisateurs des systèmes de recommandation appelés systèmes de filtrage collaboratif (FC). Notre première contribution est Hide & Share, un nouveau mécanisme de similarité, respectueux de la vie privée, pour la calcul décentralisé de graphes de K-Plus-Proches-Voisins (KPPV). C’est un mécanisme léger, conçu pour les systèmes de FC fondés sur les utilisateurs et décentralisés (ou pair-à-pair), qui se basent sur les graphes de KPPV pour fournir des recommandations. Notre seconde contribution s’applique aussi aux systèmes de FC fondés sur les utilisateurs, mais est indépendante de leur architecture. Cette contribution est double : nous évaluons d’abord l’impact d’une attaque active dite « Sybil » sur la confidentialité du profil d’intérêts d’un utilisateur cible, puis nous proposons une contre-mesure. Celle-ci est 2-step, une nouvelle mesure de similarité qui combine une bonne précision, permettant ensuite de faire de bonnes recommandations, avec une bonne résistance à l’attaque Sybil en questionRecommendation systems try to infer their users’ interests in order to suggest items relevant to them. These systems thus offer a valuable service to users in that they automatically filter non-relevant information, which avoids the nowadays common issue of information overload. This is why recommendation systems are now popular, if not pervasive in some domains such as the World Wide Web. However, an individual’s interests are personal and private data, such as one’s political or religious orientation. Therefore, recommendation systems gather private data and their widespread use calls for privacy-preserving mechanisms. In this thesis, we study the privacy of users’ interests in the family of recommendation systems called Collaborative Filtering (CF) ones. Our first contribution is Hide & Share, a novel privacy-preserving similarity mechanism for the decentralized computation of K-Nearest-Neighbor (KNN) graphs. It is a lightweight mechanism designed for decentralized (a.k.a. peer-to-peer) user-based CF systems, which rely on KNN graphs to provide recommendations. Our second contribution also applies to user-based CF systems, though it is independent of their architecture. This contribution is two-fold: first we evaluate the impact of an active Sybil attack on the privacy of a target user’s profile of interests, and second we propose a counter-measure. This counter-measure is 2-step, a novel similarity metric combining a good precision, in turn allowing for good recommendations,with high resilience to said Sybil attack
5
Strunjas, Svetlana. "Algorithms and Models for Collaborative Filtering from Large Information Corpora." University of Cincinnati / OhioLINK, 2008. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1220001182.
Full text
6
Almosallam, Ibrahim Ahmad Shang Yi. "A new adaptive framework for collaborative filtering prediction." Diss., Columbia, Mo. : University of Missouri-Columbia, 2008. http://hdl.handle.net/10355/5630.
Full textAbstract:
Thesis (M.S.)--University of Missouri-Columbia, 2008.The entire dissertation/thesis text is included in the research.pdf file; the official abstract appears in the short.pdf file (which also appears in the research.pdf); a non-technical general description, or public abstract, appears in the public.pdf file. Title from title screen of research.pdf file (viewed on August 22, 2008) Includes bibliographical references.
7
Salam, Patrous Ziad, and Safir Najafi. "Evaluating Prediction Accuracy for Collaborative Filtering Algorithms in Recommender Systems." Thesis, KTH, Skolan för datavetenskap och kommunikation (CSC), 2016. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-186456.
Full textAbstract:
Recommender systems are a relatively new technology that is commonly used by e-commerce websites and streaming services among others, to predict user opinion about products. This report studies two specific recommender algorithms, namely FunkSVD, a matrix factorization algorithm and Item-based collaborative filtering, which utilizes item similarity. This study aims to compare the prediction accuracy of the algorithms when ran on a small and a large dataset. By performing cross-validation on the algorithms, this paper seeks to obtain data that supposedly may clarify ambiguities regarding the accuracy of the algorithms. The tests yielded results which indicated that the FunkSVD algorithm may be more accurate than the Item-based collaborative filtering algorithm, but further research is required to come to a concrete conclusion.
8
NARAYANASWAMY, SHRIRAM. "A CONCEPT-BASED FRAMEWORK AND ALGORITHMS FOR RECOMMENDER SYSTEMS." University of Cincinnati / OhioLINK, 2007. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1186165016.
Full text
9
Svebrant, Henrik, and John Svanberg. "A comparative study of the conventional item-based collaborative filtering and the Slope One algorithms for recommender systems." Thesis, KTH, Skolan för datavetenskap och kommunikation (CSC), 2016. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-186449.
Full textAbstract:
Recommender systems are an important research topic in todays society as the amount of data increases across the globe. In order for commercial systems to give their users good and personalized recommendations on what data may be of interest to them in an effective manner, such a system must be able to give recommendations quickly and scale well as data increases. The purpose of this study is to evaluate two such algorithms with this in mind. The two different algorithm families tested are classified as item-based collaborative filtering but work very differently. It is therefore of interest to see how their complexities affect their performance, accuracy as well as scalability. The Slope One family is much simpler to implement and proves to be equally as efficient, if not even more efficient than the conventional item-based ones. Both families do require a precomputation stage before recommendations are possible to give, this is the stage where Slope One suffers in comparison to the conventional item-based one. The algorithms are tested using Lenskit, on data provided by GroupLens and their MovieLens project.
10
Safran, Mejdl Sultan. "EFFICIENT LEARNING-BASED RECOMMENDATION ALGORITHMS FOR TOP-N TASKS AND TOP-N WORKERS IN LARGE-SCALE CROWDSOURCING SYSTEMS." OpenSIUC, 2018. https://opensiuc.lib.siu.edu/dissertations/1511.
Full textAbstract:
A pressing need for efficient personalized recommendations has emerged in crowdsourcing systems. On the one hand, workers confront a flood of tasks, and they often spend too much time to find tasks matching their skills and interests. Thus, workers want effective recommendation of the most suitable tasks with regard to their skills and preferences. On the other hand, requesters sometimes receive results in low-quality completion since a less qualified worker may start working on a task before a better-skilled worker may get hands on. Thus, requesters want reliable recommendation of the best workers for their tasks in terms of workers' qualifications and accountability. The task and worker recommendation problems in crowdsourcing systems have brought up unique characteristics that are not present in traditional recommendation scenarios, i.e., the huge flow of tasks with short lifespans, the importance of workers' capabilities, and the quality of the completed tasks. These unique features make traditional recommendation approaches (mostly developed for e-commerce markets) no longer satisfactory for task and worker recommendation in crowdsourcing systems. In this research, we reveal our insight into the essential difference between the tasks in crowdsourcing systems and the products/items in e-commerce markets, and the difference between buyers' interests in products/items and workers' interests in tasks. Our insight inspires us to bring up categories as a key mediation mechanism between workers and tasks. We propose a two-tier data representation scheme (defining a worker-category suitability score and a worker-task attractiveness score) to support personalized task and worker recommendation. We also extend two optimization methods, namely least mean square error (LMS) and Bayesian personalized rank (BPR) in order to better fit the characteristics of task/worker recommendation in crowdsourcing systems. We then integrate the proposed representation scheme and the extended optimization methods along with the two adapted popular learning models, i.e., matrix factorization and kNN, and result in two lines of top-N recommendation algorithms for crowdsourcing systems: (1) Top-N-Tasks (TNT) recommendation algorithms for discovering the top-N most suitable tasks for a given worker, and (2) Top-N-Workers (TNW) recommendation algorithms for identifying the top-N best workers for a task requester. An extensive experimental study is conducted that validates the effectiveness and efficiency of a broad spectrum of algorithms, accompanied by our analysis and the insights gained.
11
Curnalia, James W. "The Impact of Training Epoch Size on the Accuracy of Collaborative Filtering Models in GraphChi Utilizing a Multi-Cyclic Training Regimen." Youngstown State University / OhioLINK, 2013. http://rave.ohiolink.edu/etdc/view?acc_num=ysu1370016838.
Full text
12
Williams, Alyssa. "Hybrid Recommender Systems via Spectral Learning and a Random Forest." Digital Commons @ East Tennessee State University, 2019. https://dc.etsu.edu/etd/3666.
Full textAbstract:
We demonstrate spectral learning can be combined with a random forest classifier to produce a hybrid recommender system capable of incorporating meta information. Spectral learning is supervised learning in which data is in the form of one or more networks. Responses are predicted from features obtained from the eigenvector decomposition of matrix representations of the networks. Spectral learning is based on the highest weight eigenvectors of natural Markov chain representations. A random forest is an ensemble technique for supervised learning whose internal predictive model can be interpreted as a nearest neighbor network. A hybrid recommender can be constructed by first deriving a network model from a recommender's similarity matrix then applying spectral learning techniques to produce a new network model. The response learned by the new version of the recommender can be meta information. This leads to a system capable of incorporating meta data into recommendations.
13
Kaufman, Jaime C. "A Hybrid Approach to Music Recommendation: Exploiting Collaborative Music Tags and Acoustic Features." UNF Digital Commons, 2014. http://digitalcommons.unf.edu/etd/540.
Full textAbstract:
Recommendation systems make it easier for an individual to navigate through large datasets by recommending information relevant to the user. Companies such as Facebook, LinkedIn, Twitter, Netflix, Amazon, Pandora, and others utilize these types of systems in order to increase revenue by providing personalized recommendations. Recommendation systems generally use one of the two techniques: collaborative filtering (i.e., collective intelligence) and content-based filtering.
Systems using collaborative filtering recommend items based on a community of users, their preferences, and their browsing or shopping behavior. Examples include Netflix, Amazon shopping, and Last.fm. This approach has been proven effective due to increased popularity, and its accuracy improves as its pool of users expands. However, the weakness with this approach is the Cold Start problem. It is difficult to recommend items that are either brand new or have no user activity.
Systems that use content-based filtering recommend items based on extracted information from the actual content. A popular example of this approach is Pandora Internet Radio. This approach overcomes the Cold Start problem. However, the main issue with this approach is its heavy demand on computational power. Also, the semantic meaning of an item may not be taken into account when producing recommendations.
In this thesis, a hybrid approach is proposed by utilizing the strengths of both collaborative and content-based filtering techniques. As proof-of-concept, a hybrid music recommendation system was developed and evaluated by users. The results show that this system effectively tackles the Cold Start problem and provides more variation on what is recommended.
14
Chee, Sonny Han Seng. "RecTree, a linear collaborative filtering algorithm." Thesis, National Library of Canada = Bibliothèque nationale du Canada, 2000. http://www.collectionscanada.ca/obj/s4/f2/dsk1/tape3/PQDD_0011/MQ61420.pdf.
Full text
15
U¨berall, Christian. "A dynamic multi-algorithm collaborative-filtering system." Thesis, City University London, 2012. http://openaccess.city.ac.uk/1964/.
Full textAbstract:
Nowadays users have access to an immense number of media content. They are able to consume thousands of Television (TV) channels and millions of video clips from online portals like YouTube. Due to the immense number of available content, users can have the problem to find content of interest. This problem can be solved by recommendation systems. For example, recommendation systems can be used to create recommendations which fit to the preferences of users. Recommendation systems can use two different approaches for the creation of recommendations. They can take content-based and/or collaborative-filtering techniques into account. Content-based filtering techniques use information, the so-called metadata, that describe the content in more detail. Collaborative-filtering techniques calculate similarities e.g., between users. All users are included in a dataset, the so-called community. Generally the number of user profiles within the community is quite large. Examples of such huge communities are Amazon, Netflix, MovieLens, and LastFM. The community which includes the user profiles is used to create a user-item matrix. This user-item matrix contains the preferences from users on items e.g., movies, genres, book titles, and so forth. The quality of the recommendations depends on the accuracy of the predictions. As mentioned above, collaborative-filtering techniques calculate similarities e.g., between users. These similarities can be used to calculate predictions for an entry within the user-item matrix. If the predictions are close or equal to the preferences of a user, the used collaborative-filtering technique predicts accurately. Generally recommendation systems only use one single collaborative-filtering algorithm for the similarity calculation. The research work of this thesis proves that a dynamic selection of the most accurate filtering algorithm by considering more algorithms is able to increase the accuracy of the predictions significantly. In order to increase the accuracy of predictions, this thesis presents a dynamic multi-algorithm collaborative-filtering system which creates recommendations for video content, such as movies or genres. This system is able to find the most accurate filtering algorithm by considering the k-nearest neighbours. These neighbours are selected by identifying the most similar users or items e.g., movies. Besides the dynamic selection, this thesis presents newly developed collaborative-filtering algorithms which are able to overcome researched weaknesses of state-of-the-art algorithms. The evaluation of the proposed system considers a huge dataset from MovieLens and a small dataset from an undertaken survey. The consideration of a huge and a small dataset shall prove that the system can be used in both cases. The results of this thesis show that the proposed system is able to decrease the error rate significantly compared to existing approaches.
16
Tetler, William G. (William Gore). "A collaborative filtering prediction algorithm for ClassRank subject recommendations." Thesis, Massachusetts Institute of Technology, 2007. http://hdl.handle.net/1721.1/46521.
Full textAbstract:
Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, February 2008.Includes bibliographical references (p. 51).
Undergraduate students at M.I.T. typically utilize three resources when selecting subjects: course specific evaluations, faculty advisors, and peers. While these resources have distinct advantages, they are all limited in scope. The ClassRank web application has been developed to bridge the gap between these resources by providing a simple institute-wide system for undergraduate students to evaluate and rate subjects. The application also provides a solid platform to build new tools utilizing subject evaluation data. To extend the initial core functionality of the ClassRank system, a rating-based subject recommendation algorithm was added to offer students an unbiased perspective on potential subjects of interest. Developed as a Ruby on Rails plugin and then integrated into ClassRank, the recommendation algorithm analyzes subject ratings and provides personalized suggestions to students about subjects that would likely fit their interests and educational goals. The ClassRank web application and recommendation algorithm will provide the M.I.T. undergraduate student body with a unique and invaluable resource for subject selection.
by William G. Tetler.
M.Eng.
17
Pascoal, Luiz Mário Lustosa. "Um método social-evolucionário para geração de rankings que apoiem a recomendação de eventos." Universidade Federal de Goiás, 2014. http://repositorio.bc.ufg.br/tede/handle/tede/4345.
Full textAbstract:
Submitted by Erika Demachki (erikademachki@gmail.com) on 2015-03-24T21:17:09Z
No. of bitstreams: 3
Dissertação - Luiz Mario Lustosa Pascoal - 2014.pdf: 7280181 bytes, checksum: 68a6ac0602e3e51f6e6952bbd6916150 (MD5)
FunctionApproximator.zip: 2288624 bytes, checksum: 178c2e6a0b080b3d0548836974016236 (MD5)
license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5)Approved for entry into archive by Erika Demachki (erikademachki@gmail.com) on 2015-03-24T21:19:16Z (GMT) No. of bitstreams: 3 Dissertação - Luiz Mario Lustosa Pascoal - 2014.pdf: 7280181 bytes, checksum: 68a6ac0602e3e51f6e6952bbd6916150 (MD5) FunctionApproximator.zip: 2288624 bytes, checksum: 178c2e6a0b080b3d0548836974016236 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5)
Made available in DSpace on 2015-03-24T21:19:16Z (GMT). No. of bitstreams: 3 Dissertação - Luiz Mario Lustosa Pascoal - 2014.pdf: 7280181 bytes, checksum: 68a6ac0602e3e51f6e6952bbd6916150 (MD5) FunctionApproximator.zip: 2288624 bytes, checksum: 178c2e6a0b080b3d0548836974016236 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2014-08-22
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES
With the development of web 2.0, social networks have achieved great space on the internet, with that many users provide information and interests about themselves. There are expert systems that make use of the user’s interests to recommend different products, these systems are known as Recommender Systems. One of the main techniques of a Recommender Systems is the Collaborative Filtering (User-based) which recommends products to users based on what other similar people liked in the past. Therefore, this work presents model approximation of functions that generates rankings, that through a Genetic Algorithm, is able to learn an approximation function composed by different social variables, customized for each Facebook user. The learned function must be able to reproduce a ranking of people (friends) originally created with user’s information, that apply some influence in the user’s decision. As a case study, this work discusses the context of events through information regarding the frequency of participation of some users at several distinct events. Two different approaches on learning and applying the approximation function have been developed. The first approach provides a general model that learns a function in advance and then applies it in a set of test data and the second approach presents an specialist model that learns a specific function for each test scenario. Two proposals for evaluating the ordering created by the learned function, called objective functions A and B, where the results for both objective functions show that it is possible to obtain good solutions with the generalist and the specialist approaches of the proposed method.
Com o desenvolvimento da Web 2.0, as redes sociais têm conquistado grande espaço na internet, com isso muitos usuários acabam fornecendo diversas informações e interesses sobre si mesmos. Existem sistemas especialistas que fazem uso dos interesses do usuário para recomendar diferentes produtos, esses sistemas são conhecidos como Sistemas de Recomendação. Uma das principais técnicas de um Sistema de Recomendação é a Filtragem Colaborativa (User-based) que recomenda produtos para seus usuários baseados no que outras pessoas similares à ele tenham gostado no passado. Portanto, este trabalho apresenta um modelo de aproximação de funções geradora de rankings que, através de um Algoritmo Genético, é capaz de aprender uma função de aproximação composta por diferentes atributos sociais, personalizada para cada usuário do Facebook. A função aprendida deve ser capaz de reproduzir um ranking de pessoas (amigos) criado originalmente com informações do usuário, que exercem certa influência na decisão do usuário. Como estudo de caso, esse trabalho aborda o contexto de eventos através de informações com relação a frequência de participação de alguns usuários em vários eventos distintos. Foram desenvolvidas duas abordagens distintas para aprendizagem e aplicação da função de aproximação. A primeira abordagem apresenta um modelo generalista, que previamente aprende uma função e em seguida a aplica em um conjunto de dados de testes e a segunda abordagem apresenta um modelo especialista, que aprende uma função específica para cada cenário de teste. Também foram apresentadas duas propostas para avaliação da ordenação criada pela função aprendida, denominadas funções objetivo A e B, onde os resultados para ambas as funções objetivo A e B mostram que é possível obter boas soluções com as abordagens generalista e especialista do método proposto.
18
Jiahui, Yu. "Research on collaborative filtering algorithm based on knowledge graph and long tail." Thesis, Blekinge Tekniska Högskola, Institutionen för datavetenskap, 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:bth-18828.
Full textAbstract:
Background: With the popularization of the Internet and the development of information technology, the network information data has shown an explosive growth, and the problem of information overload [1] has been highlighted. In order to help users, find the information they are interested in from a large amount of information, and help information producers to let their own information be concerned by the majority of users, the recommendation system came into being. Objectives: However, the sparseness problem, the neglect of semantic information, and the failure to consider the coverage rate faced by the traditional recommendation system limit the effect of the recommendation system to some extent. So in this paper I want to deal with these problems. Methods: This paper improves the performance of the recommendation system by constructing a knowledge graph in the domain and using knowledge embedding technology (openKE), combined with the collaborative filtering algorithm based on the long tail theory. And I use 3 experiments to verify this proposed approach’s performance of recommendation and the ability to dig the long tail information, I compared it with some other collaborative filtering algorithms. Results: The results show that the proposed approach improves the precision, recall and coverage and has a better ability to mine the long tail information. Conclusion: The proposed method improves the recommended performance by reducing the sparsity of the matrix and mining the semantic information between the items. At the same time, the long tail theory is considered, so that users can be recommended to more items that may be of interest.
19
Joshi, Bikash. "Algorithmes d'apprentissage pour les grandes masses de données : Application à la classification multi-classes et à l'optimisation distribuée asynchrone." Thesis, Université Grenoble Alpes (ComUE), 2017. http://www.theses.fr/2017GREAM046/document.
Full textAbstract:
L'objectif de cette thèse est de développer des algorithmes d'apprentissage adaptés aux grandes masses de données. Dans un premier temps, nous considérons le problème de la classification avec un grand nombre de classes. Afin d'obtenir un algorithme adapté à la grande dimension, nous proposons un algorithme qui transforme le problème multi-classes en un problème de classification binaire que nous sous-échantillonnons de manière drastique. Afin de valider cette méthode, nous fournissons une analyse théorique et expérimentale détaillée.Dans la seconde partie, nous approchons le problème de l'apprentissage sur données distribuées en introduisant un cadre asynchrone pour le traitement des données. Nous appliquons ce cadre à deux applications phares : la factorisation de matrice pour les systèmes de recommandation en grande dimension et la classification binaireThis thesis focuses on developing scalable algorithms for large scale machine learning. In this work, we present two perspectives to handle large data. First, we consider the problem of large-scale multiclass classification. We introduce the task of multiclass classification and the challenge of classifying with a large number of classes. To alleviate these challenges, we propose an algorithm which reduces the original multiclass problem to an equivalent binary one. Based on this reduction technique, we introduce a scalable method to tackle the multiclass classification problem for very large number of classes and perform detailed theoretical and empirical analyses.In the second part, we discuss the problem of distributed machine learning. In this domain, we introduce an asynchronous framework for performing distributed optimization. We present application of the proposed asynchronous framework on two popular domains: matrix factorization for large-scale recommender systems and large-scale binary classification. In the case of matrix factorization, we perform Stochastic Gradient Descent (SGD) in an asynchronous distributed manner. Whereas, in the case of large-scale binary classification we use a variant of SGD which uses variance reduction technique, SVRG as our optimization algorithm
20
Ozturk, Gizem. "A Hybrid Veideo Recommendation System Based On A Graph Based Algorithm." Master's thesis, METU, 2010. http://etd.lib.metu.edu.tr/upload/12612624/index.pdf.
Full textAbstract:
This thesis proposes the design, development and evaluation of a hybrid video
recommendation system. The proposed hybrid video recommendation system is based
on a graph algorithm called Adsorption. Adsorption is a collaborative filtering algorithm
in which relations between users are used to make recommendations. Adsorption is used
to generate the base recommendation list. In order to overcome the problems that occur
in pure collaborative system, content based filtering is injected. Content based filtering
uses the idea of suggesting similar items that matches user preferences. In order to use
content based filtering, first, the base recommendation list is updated by removing weak
recommendations. Following this, item similarities of the remaining list are calculated
and new items are inserted to form the final recommendations. Thus, collaborative
recommendations are empowered considering item similarities. Therefore, the
developed hybrid system combines both collaborative and content based approaches to
produce more effective suggestions.
21
Bookwala, Avinash Turab. "Combined map personalisation algorithm for delivering preferred spatial features in a map to everyday mobile device users." AUT University, 2009. http://hdl.handle.net/10292/920.
Full textAbstract:
In this thesis, we present an innovative and novel approach to personalise maps/geo-spatial services for mobile users. With the proposed map personalisation approach, only relevant data will be extracted from detailed maps/geo-spatial services on the fly, based on a user’s current location, preferences and requirements. This would result in dramatic improvements in the legibility of maps on mobile device screens, as well as significant reductions in the amount of data being transmitted; which, in turn, would reduce the download time and cost of transferring the required geo-spatial data across mobile networks. Furthermore, the proposed map personalisation approach has been implemented into a working system, based on a four-tier client server architecture, wherein fully detailed maps/services are stored on the server, and upon a user’s request personalised maps/services, extracted from the fully detailed maps/services based on the user’s current location, preferences, are sent to the user’s mobile device through mobile networks. By using open and standard system development tools, our system is open to everyday mobile devices rather than smart phones and Personal Digital Assistants (PDA) only, as is prevalent in most current map personalisation systems. The proposed map personalisation approach combines content-based information filtering and collaborative information filtering techniques into an algorithmic solution, wherein content-based information filtering is used for regular users having a user profile stored on the system, and collaborative information filtering is used for new/occasional users having no user profile stored on the system. Maps/geo-spatial services are personalised for regular users by analysing the user’s spatial feature preferences automatically collected and stored in their user profile from previous usages, whereas, map personalisation for new/occasional users is achieved through analysing the spatial feature preferences of like-minded users in the system in order to make an inference for the target user. Furthermore, with the use of association rule mining, an advanced inference technique, the spatial features retrieved for new/occasional users through collaborative filtering can be attained. The selection of spatial features through association rule mining is achieved by finding interesting and similar patterns in the spatial features most commonly retrieved by different user groups, based on their past transactions or usage sessions with the system.
22
Ozbal, Gozde. "A Content Boosted Collaborative Filtering Approach For Movie Recommendation Based On Local &." Master's thesis, METU, 2009. http://etd.lib.metu.edu.tr/upload/12610984/index.pdf.
Full textAbstract:
Recently, it has become more and more difficult for the existing web based systems
to locate or retrieve any kind of relevant information, due to the rapid growth of the
World Wide Web (WWW) in terms of the information space and the amount of the
users in that space. However, in today's world, many systems and approaches make it possible for the users to be guided by the recommendations that they provide about new items such as articles, news, books, music, and movies. However, a lot of traditional recommender systems result in failure when the data to be used throughout the recommendation process is sparse. In another sense, when there exists an inadequate number of items or users in the system, unsuccessful recommendations are produced. Within this thesis work, ReMovender, a web based movie recommendation system, which uses a content boosted collaborative filtering approach, will be presented. ReMovender combines the local/global similarity and missing data prediction v techniques in order to handle the previously mentioned sparseness problem effectively. Besides, by putting the content information of the movies into consideration during the item similarity calculations, the goal of making more successful and realistic predictions is achieved.
23
ROSSETTI, MARCO. "Advancing Recommender Systems from the Algorithm, Interface and Methodological Perspective." Doctoral thesis, Università degli Studi di Milano-Bicocca, 2015. http://hdl.handle.net/10281/70560.
Full textAbstract:
I sistemi di raccomandazione sono componenti software che aiutano gli utenti a trovare quello che stanno cercando. I sistemi di raccomandazione sono stati applicati a diverse aree, dal commercio elettronico alle notizie, dalla musica al turismo, sfruttando tutte le informazioni disponibili per imparare le preferenze dell’utente e fornire raccomandazioni utili.
La vasta area dei sistemi di raccomandazione riguarda molte tematiche che richiedono una conoscenza profonda e grandi sforzi di ricerca. In particolare, tre aspetti principali sono: algoritmi, ossia i componenti intelligenti che elaborano le raccomandazioni; interfacce, ossia gli strumenti che permettono di mostrare le raccomandazioni agli utenti; valutazione, ossia le metodologie per validare l’efficacia dei sistemi di raccomandazione.
In questa dissertazione ci focalizziamo su questi aspetti guidati da tre considerazioni. Primo, il contenuto testuale relativo agli item e ai rating può essere sfruttato per migliorare diversi aspetti, come elaborare raccomandazioni, fornire spiegazioni e comprendere i gusti degli utenti e le potenzialità degli item. Secondo, il tempo nei sistemi di raccomandazione dovrebbe essere considerato in quanto ha una grande influenza sulla popolarità e sui gusti. Terzo, i protocolli di valutazione offline non sono completamente convincenti, in quanto si basano su statistiche di accuratezza che non sempre rispecchiano le reali preferenze dell’utente.
Date le motivazioni citate, vengono forniti sei contributi divisi tra l’integrazione di concetti e tempo nei sistemi di raccomandazione, l’applicazione del topic model per analizzare recensioni e spiegare fattori latenti, e la validazione delle misure di valutazione offline.Recommender systems are software components that assist users in finding what they are looking for. They have been applied to all kinds of domains, from ecommerce to news, from music to tourism, exploiting all the information available in order to learn user's preferences and to provide useful recommendations. The broad area of recommender systems has many topics that require a deep understanding and great research efforts. In particular, three main aspects are: algorithms, which are the hidden intelligent components that compute recommendations; interfaces, which are the way in which recommendations are shown to the user; evaluation, which is the methodology to assess the effectiveness of a recommender system. In this dissertation we focus on these aspects guided by three considerations. First, textual content related to items and ratings can be exploited in order to improve several aspects, such as to compute recommendations, provide explanations, understand user's tastes and item's capabilities. Second, time in recommender systems should be considered as it has a great influence on popularity and tastes. Third, offline evaluation protocols are not fully convincing, as they are based on accuracy statistics that do not always reflect real user's preferences. Following these motivations six contributions have been delivered, broadly divided in the integration of concepts and time in recommender systems, the application of the topic model to analyze user reviews and to explain latent factors, and the validation of offline recommendation accuracy measurements.
24
Ye, Brian, and Benny Tieu. "Implementation and Evaluation of a Recommender System Based on the Slope One and the Weighted Slope One Algorithm." Thesis, KTH, Skolan för datavetenskap och kommunikation (CSC), 2015. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-166438.
Full textAbstract:
Recommender systems are used on many different websites today and are mechanisms that are supposed to accurately give personalized recommendations of items to a set of different users. An item can for example be movies on Netflix. The purpose of this paper is to implement an algorithm that fulfills five stated goals of the implementation. The goals are as followed: the algorithm should be easy to implement, be effective on query time, accurate on recommendations, put little expectations on users and alternations of algorithm should not have to be changed comprehensively. Slope One is a simplified version of linear regression and can be used to recommend items. By using the Netflix Prize data set from 2009 and the Root-Mean-Square-Error (RMSE) as an evaluator, Slope One generates an accuracy of 1.007 units. The Weighted Slope One, which takes the relevancy of items into the calculation, generates an accuracy of 0.990 units. Adding Weighted Slope One to the Slope One implementation can be done without changing the fundamentals of the Slope One algorithm. It is nearly instantaneous to generate a recommendation of a movie with regular Slope One and Weighted Slope One. However, a precomputing stage is needed for the mechanism. In order to receive a recommendation of the implementation in this paper, the user must at least have rated two items.Rekommendationssystem används idag på många olika hemsidor, och är en mekanism som har syftet att, med noggrannhet, ge en personlig rekommendation av objekt till en mängd olika användare. Ett objekt kan exempelvis vara en film från Netflix. Syftet med denna rapport är att implementera en algoritm som uppfyller fem olika implementationsmål. Målen är enligt följande: algoritmen ska vara enkel att implementera, ha en effektiv tid på dataförfrågan, ge noggranna rekommendationer, sätta låga förväntningar hos användaren samt ska algoritmen inte behöva omfattande förändring vid alternering. Slope One är en förenklad version av linjär regression, och kan även användas till att rekommendera objekt. Genom att använda datamängden från Netflix Prize från 2009 och måttet Root-Mean-Square-Error (RMSE) som en utvärderare, kan Slope One generera en precision på 1.007 enheter. Den viktade Slope One, som tar hänsyn till varje föremåls relevans, genererar en precision på 0.990 enheter. När dessa två algoritmer kombineras, behövs inte större fundamentala ändringar i implementationen av Slope One. En rekommendation av något objekt kan genereras omedelbart med någon av de två algoritmerna, dock krävs det en förberäkningsfas i mekanismen. För att få en rekommendation av implementationen i denna rapport, måste användaren åtminstone ha värderat två objekt.
25
Bodeček, Miroslav. "Algoritmus pro cílené doporučování produktů." Master's thesis, Vysoké učení technické v Brně. Fakulta informačních technologií, 2011. http://www.nusl.cz/ntk/nusl-412860.
Full textAbstract:
The goal of this project is to explore the problem of product recommendations in the area of e-commerce and to evaluate known techniques, design product recommendation system for an existing e-commerce site, implement it and test it. This report introduces the problem, briefly examines current state of affairs in this area and defines requirements for a product recommendation module. The concept of data mining in general is introduced. The report proceeds to present detailed design corresponding to defined requirements and summarizes data gathered during testing phase. It concludes with evaluation and with discussion of the remaining goals for this thesis.
26
Paraschakis, Dimitris. "Algorithmic and Ethical Aspects of Recommender Systems in e-Commerce." Licentiate thesis, Malmö universitet, Fakulteten för teknik och samhälle (TS), 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:mau:diva-7792.
Full textAbstract:
Recommender systems have become an integral part of virtually every e-commerce application on the web. These systems enable users to quickly discover relevant products, at the same time increasing business value. Over the past decades, recommender systems have been modeled using numerous machine learning techniques. However, the adoptability of these models by commercial applications remains unclear. We assess the receptiveness of the industrial sector to algorithmic contributions of the research community by surveying more than 30 e-commerce platforms, and experimenting with various recommenders on proprietary e-commerce datasets. Another overlooked but important factor that complicates the design and use of recommender systems is their ethical implications. We identify and summarize these issues in our ethical recommendation framework, which also enables users to control sensitive moral aspects of recommendations via the proposed “ethical toolbox”. The feasibility of this tool is supported by the results of our user study. Because of moral implications associated with user profiling, we investigate algorithms capable of generating user-agnostic recommendations. We propose an ensemble learning scheme based on Thompson Sampling bandit policy, which models arms as base recommendation functions. We show how to adapt this algorithm to realistic situations when neither arm availability nor reward stationarity is guaranteed.
27
Aleixo, Everton Lima. "Item-based-adp: análise e melhoramento do algoritmo de filtragem colaborativa item-based." Universidade Federal de Goiás, 2014. http://repositorio.bc.ufg.br/tede/handle/tede/4133.
Full textAbstract:
Submitted by Erika Demachki (erikademachki@gmail.com) on 2015-02-06T20:35:15Z
No. of bitstreams: 2
Dissertação - Everton Lima Aleixo - 2014.pdf: 2375638 bytes, checksum: accbd56745e040e23362d951a1336538 (MD5)
license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5)Approved for entry into archive by Erika Demachki (erikademachki@gmail.com) on 2015-02-06T20:35:41Z (GMT) No. of bitstreams: 2 Dissertação - Everton Lima Aleixo - 2014.pdf: 2375638 bytes, checksum: accbd56745e040e23362d951a1336538 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5)
Made available in DSpace on 2015-02-06T20:35:41Z (GMT). No. of bitstreams: 2 Dissertação - Everton Lima Aleixo - 2014.pdf: 2375638 bytes, checksum: accbd56745e040e23362d951a1336538 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2014-09-02
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES
Memory-based algorithms are the most popular among the collaborative filtering algorithms. They use as input a table containing ratings given by users to items, known as the rating matrix. They predict the rating given by user a to an item i by computing similarities of the ratings among users or similarities of the ratings among items. In the first case Memory-Based algorithms are classified as User-based algorithms and in the second one they are labeled as Item-based algorithms. The prediction is computed using the ratings of k most similar users (or items), also know as neighbors. Memory-based algorithms are simple to understand and to program, usually provide accurate recommendation and are less sensible to data change. However, to obtain the most similar neighbors for a prediction they have to process all the data which is a serious scalability problem. Also they are sensitive to the sparsity of the input. In this work we propose an efficient and effective Item-Based that aims at diminishing the sensibility of the Memory-Based approach to both problems stated above. The algorithm is faster (almost 50%) than the traditional Item-Based algorithm while maintaining the same level of accuracy. However, in environments that have much data to predict and few to train the algorithm, the accuracy of the proposed algorithm surpass significantly that of the traditional Item-based algorithms. Our approach can also be easily adapted to be used as User-based algorithms.
Algoritmos baseados em memória são os mais populares entre os algoritmos de filtragem colaborativa. Eles usam como entrada uma tabela contendo as avaliações feitas pelos usuários aos itens, conhecida como matriz de avaliações. Eles predizem a avaliação dada por um usuário a a um item i, computando a similaridade de avaliações entre a e outros usuários ou entre i e outros itens. No primeiro caso, os algoritmos baseados em memória são classificados como algoritmos baseados em usuários (User-based) e no segundo caso são rotulados como algoritmos baseados em itens (Item-Based). A predição é computada usando as avaliações dos k usuários (ou itens) mais similares, também conhecidos como vizinhos. Algoritmos baseados em memória são simples de entender e implementar. Normalmente produzem boas recomendações e são menos sensíveis a mudança nos dados. Entretanto, para obter os vizinhos mais similares para a predição, eles necessitam processar todos os dados da matriz, o que é um sério problema de escalabilidade. Eles também são sensíveis a densidade dos dados. Neste trabalho, nós propomos um algoritmo eficiente e eficaz baseado em itens que visa diminuir a sensibilidade dos algoritmos baseados em memória para ambos os problemas acima referidos. Esse algoritmo é mais rápido (quase 50%) do que o algoritmo baseado em itens tradicional, mantendo o mesmo nível de acurácia. Entretanto, em ambientes onde existem muitos dados para predizer e poucos para treinar o algoritmo, a acurácia do algoritmo proposto supera significativamente a do algoritmo tradicional baseado em itens. Nossa abordagem pode ainda ser facilmente adaptada para ser utilizada como o algoritmo baseado em usuários.
28
Alkilicgil, Erdem. "User Modeling In Mobile Environment." Master's thesis, METU, 2005. http://etd.lib.metu.edu.tr/upload/12606852/index.pdf.
Full textAbstract:
The popularity of e-commerce sites and applications that use recommendations and
user modeling is increased recently. The development and contest in tourism calls
attention of large-scale IT companies. These companies have started to work on
recommendation systems and user modeling on tourism sector. Some of the
clustering methodologies, neighboring methods and machine learning algorithms are
commenced to use for making predictions about tourist&rsquos interests while he/she is traveling around the city. Recommendation ability is the most interesting thing for a tourist guide application. Recommender systems are composed of two main approaches, collaborative and content-based filtering. Collaborative filtering algorithms look for people that have similar interests and properties, while contentbased filtering methods pay attention to sole user&rsquo
s interests and properties to make recommendations. Both of the approaches have advantages and disadvantages, for that reason sometimes these two approaches are used together. Chosen method directly affects the recommendation quality, so advantages and disadvantages of both methods will be examined carefully. Recommendation of locations or services can be seen as a classification problem. Artificial intelligent systems like neural networks, genetic algorithms, particle swarm optimization algorithms, artificial immune systems are inspired from natural life and can be used as classifier systems. Artificial immune system, inspired from human immune system, has ability to classify huge numbers of different patterns. In this paper ESGuide, a tourist guide application that uses artificial immune system is examined. ESGuide application is a client-server application that helps tourists while they are traveling around the city. ESGuide has two components: Map agent and recommender agent. Map agent helps the tourist while he/she interacts with the city map. Tourist should rate the locations and items while traveling. Due to these ratings and client-server interaction, recommender agent tries to predict user interested places and items. Tourist has a chance to state if he/she likes the recommendation or not. If the tourist does not like the recommendation, new recommendation set is created and presented to the user.
29
Ma, Chih-Chao, and 馬智釗. "Large-scale Collaborative Filtering Algorithms." Thesis, 2008. http://ndltd.ncl.edu.tw/handle/90098768318556866413.
Full textAbstract:
碩士國立臺灣大學
資訊工程學研究所
96
As the market of electronic commerce grows explosively, it is important to provide customized suggestions for various consumers. Collaborative filtering is an important technique which models and analyzes the preferences of customers, and gives suitable advices. In this thesis, we study large-scale collaborative filtering algorithms to process huge data sets in acceptable time. We use the well-known Singular Value Decomposition as the basis of our algorithms, and propose some improvements. We also discuss post-processing methods. We participate at the competition of Netflix Prize, a contest of predicting movie preferences, and achieve good results.
30
Ma, Chih-Chao. "Large-scale Collaborative Filtering Algorithms." 2008. http://www.cetd.com.tw/ec/thesisdetail.aspx?etdun=U0001-2506200816405300.
Full text
31
Dileep, K., and C. M. Rao. "Analysis of collaborative filtering algorithms." Thesis, 2014. http://ethesis.nitrkl.ac.in/5580/1/E-THESIS_44.pdf.
Full textAbstract:
Recommender System is a subclass of information filtering system which predicts the rating given to an item by any user. Collaborative filtering is a key technique in recommender systems. This technique predicts the user rating of an item by collaboration of other users who have similar interests with this user. Collaborative filtering approaches can be categorized as Memory based, Model-based and Hybrid approaches. Memory-based approach can be further classified as Item-based and User-based recommendations. Pearson correlation scheme belongs to user-based scheme and Slope one family of algorithms belong to item-based scheme. Slope one family consists of Normal, Weighted and Bipolar slope one algorithms. Algorithms belonging to model-based approach are Singular value decomposition, Regularized Singular value decomposition and Probabilistic Matrix Factorization. In hybrid approach combination of memory-based and model-based approaches are used for making recommendations. In this thesis we made an attempt to analyze various algorithms in Memory-based and Model-based approaches. In model based algorithms, we analyzed Singular Value Decomposition (SVD) and Regularized Singular Value Decomposition (RSVD). By taking three different dataset sizes, we observed that RSVD outperforms SVD for all three dataset sizes. In memory based algorithms, we analyzed Pearson correlation scheme which takes the correlation between user vectors as similarity measure and Slope one family of algorithms. In slope one algorithms, we proposed an improvement to the existing scheme for determining Threshold value of Bipolar slope one algorithm. We used median and average of min-max ratings which outperforms the existing user average scheme. Finally, we made an analysis of all these algorithms and concluded that RSVD outperforms rest of the algorithms in terms of accuracy of predictions.
32
Lin, Chin-Ta, and 林敬達. "Implementation and Application of Collaborative Filtering Algorithms." Thesis, 2007. http://ndltd.ncl.edu.tw/handle/94160543968912276482.
Full textAbstract:
碩士淡江大學
資訊工程學系碩士班
95
The rise of the network job bank, not only broke the restriction of the timespace, and make the enterprise made use of the cheap cost of the network, let the tradition wanted spend the expense that a money publishes the hiring apocalypse and prints to make the DM(the advertisement handbill),the manufacture television advertisement … ...etc., all transfer the internet. Past biggest problem is not an information to obtain today, but the information overloads. This research purpose builds up a set of" recommend system of collaborative filtering ", providing the youth student that will soon graduate, at the devotion occupation place the direction and suggestion needed, hope by this mechanism recommend the studentses to lack the interested in work job, with decrease is planless to browse and waste time on the network job bank, then can find out its most fit work.
33
Hsiao, Yi-Ting, and 蕭伊廷. "Big Data Analysis and Performance Evaluation Collaborative Filtering Algorithms." Thesis, 2017. http://ndltd.ncl.edu.tw/handle/8mvdy5.
Full textAbstract:
碩士東海大學
資訊工程學系
105
In order to satisfy various demands of different Big Data applications, many rec- ommendation algorithms have been proposed, in which the collaborative filtering approach has been widely adopted. In certain situations, the Pearson correlation coefficient used in the collaborative filtering algorithm will be incorrect. In this thesis, we propose to use the Normal Recovery Similarity Measure to modify the similarity value in order to reduce the error of collaborative filtering recommen- dation algorithm. We implemnet the proposed collaborative filtering system on a stand alone PC and a cloud computing environment with 3, 6 and 9 nodes. The execution time and performance of the proposed system are measured and analyzed. From the experimental results, we find that, by cloud computing, per- formance of the proposed collaborative filtering system can be effectively enhanced.
34
Ting, Wu-Chou, and 丁吳稠. "An E-Commerce Recommendation Platform Using Collaborative and Content-based Filtering Algorithms." Thesis, 2015. http://ndltd.ncl.edu.tw/handle/2ynv7z.
Full textAbstract:
碩士國立臺北科技大學
電資碩士班
103
Recommendation system, RS, which be exist in order to dig out hidden valuable information from mega data, is used to improve the making-decision for consumers. With emerging development of Internet and E-Commerce, the quickly exchange among information and surprising growth by unlimited way will result in the appearance of Information Overload. It’s a huge challenge for the e-commerce industry how to seek out business opportunity from Big Data and then assists consumers to find information which is actually need to them. Recommendation system is a kind of service system which offers suggestion item by based on users’ preference, and has been the application of Collaborative Filtering and Content-based recommendation in extensively way. This study will using Collaborative and Content-based Filtering Algorithms and add 2 diversity factors. As this experiment told, the study that mention the method of Collaborative Filtering and Content-based recommendation can reinforce the accuracy and diversity of recommendation item and be the better solution to comparable with other related studying research. Moreover, by classify the consumers’ history to adjust algorithmic method, which fits well in users, can be a great solution to the problem of Collaborative Filtering, with problem on Cold-Start and Scalability, and can cause this system more flexible and modulated.
35
SINGH, YOGENDRA. "A PERSONALIZED HYBRID MOVIE RECOMMENDATION SYSTEM FOR USERS." Thesis, 2016. http://dspace.dtu.ac.in:8080/jspui/handle/repository/15145.
Full textAbstract:
We describe a rating logical thinking approach to incorporating matter user reviews into Collaborative Filtering (CF) algorithms. The main motive of our approach is to use user preferences which is expressed in movie reviews and then convert such user’s preferences into some rating that may be understood by existing CF algorithms. The linguistics score of subjective sentence is fetched from SentiWordNet Library to calculate their sentiments as +ve, -ve or neutral based on the textual review. We’ve used SentiWordNet library as a dataset with two completely different approaches of alternatives comprising of adverbs and verbs, adjectives and n-gram feature extraction. We have a tendency to conjointly used our SentiWordNet library to figure the document level sentiment for every movie reviewed and compared its label with results obtained victimization Alchemy API. We conjointly developed and evaluated a model of the planned framework. Preliminary results valid the effectiveness of varied tasks within the planned framework, and recommend that the framework doesn't admit an oversized coaching corpus to operate. Additional development of our rating logical thinking framework is in progress. A comprehensive analysis of the framework are administered and reported during a follow-up article.
36
Vaz, Rui Fernando Martins. "Estratégias de filtragem Anti-Spam baseadas em técnicas de computação evolucionária." Master's thesis, 2012. http://hdl.handle.net/1822/19599.
Full textAbstract:
Dissertação de mestrado integrado em Engenharia de ComunicaçõesO serviço de correio eletrónico é atualmente um serviço de comunicação essencial, que assume uma crescente importância na sociedade atual. No entanto, apesar dos vários esforços concentrados contra o correio eletrónico (email) não solicitado, designado também por spam, este continua a ser ainda um problema inerente a este serviço. No quadro das soluções tecnológicas, os métodos de filtragem baseados no conteúdo das mensagens de email, que utilizam técnicas de data mining, são atualmente os mais populares e amplamente utilizados para combater este problema. No âmbito destes métodos, a seleção de atributos que melhor caraterizam as mensagens de spam (e.g., palavras mais correlacionadas com mensagens de spam), constitui um passo importante no desenvolvimento de filtros mais assertivos. Nesse sentido, é efetuado neste trabalho um estudo empírico da introdução de técnicas de computação evolucionária de seleção de atributos no contexto da filtragem anti-spam. De forma a avaliar o método proposto foram desenvolvidos diversos filtros anti- spam que implementam, usando estratégias diferentes, técnicas de computação evolucionária de seleção de atributos. Uma das estratégias desenvolvida segue uma abordagem colaborativa que permite a troca de atributos relevantes entre filtros locais. O desempenho dos filtros anti-spam que utilizam técnicas de computação evolucionária de seleção de atributos são analisados. Posteriormente o desempenho do filltro colaborativo e comparado com um filtro padrão que utiliza apenas um método de seleção de atributos baseado num critério de informação.
Nowadays electronic mail (email) service assumes an increasing importance in modern society and is considered an essential communication service. However, despite the several efforts made against the unsolicited email (also known as spam), it remains an inherent problem which affects this service. Within the existing technological solutions, Content-Based Filtering (CBF) methods, that use data mining techniques, are currently the most popular approaches to solve this issue. Feature selection techniques are essential in CBF methods. These techniques allow the selection of a reduced set of relevant attributes (e.g., words correlated with spam messages) that provides essential information to enhance the accuracy of anti-spam filters. Hence, in this work we perform an empirical study concerning the introduction of evolutionary computation techniques for feature selection in the scope of anti-spam filtering. In order to evaluate the proposed method several anti-spam filters were developed. These filters implement, through different strategies, evolutionary computation techniques for feature selection. One of these strategies follows a collaborative approach which enables the exchange of relevant attributes between local filters. The performances of the developed filters that implement evolutionary computation techniques are evaluated. Afterwards, the performance of the collaborative filter is compared to a standard filter which uses a feature selection method based on an information criterion.
Fundação para a Ciência e a Tecnologia (FCT) - Projecto de R&D PTDC/EIA/64541/2006 - "SPAM Telescope Miner: detecção a nível mundial de correio electrónico não solicitado via técnicas de data mining"
37
Κουνέλη, Μαριάννα. "Ανάπτυξη συστήματος συστάσεων συνεργατικής διήθησης με χρήση ιεραρχικών αλγορίθμων κατάταξης." Thesis, 2012. http://hdl.handle.net/10889/5826.
Full textAbstract:
Σκοπός της παρούσας διπλωματικής διατριβής είναι η μελέτη και ανάπτυξη ενός νέου αλγοριθμικού πλαισίου Συνεργατικής Διήθησης(CF) για την παραγωγή συστάσεων. Η μέθοδος που προτείνουμε, βασίζεται στην εκμετάλλευση της ιεραρχικής διάρθρωσης του χώρου αντικειμένων και πατά διαισθητικά στην ιδιότητα της ``Σχεδόν Πλήρης Αναλυσιμότητας'' (NCD) η οποία είναι συνυφασμένη με τη δομή της πλειοψηφίας των ιεραρχικών συστημάτων.
Η Συνεργατική Διήθηση αποτελεί ίσως την πιο πετυχημένη οικογένεια τεχνικών για την παραγωγή συστάσεων. Η μεγάλη απήχησή της στο διαδίκτυο αλλά και η ευρεία εφαρμογή της σε σημαντικά εμπορικά περιβάλλοντα, έχουν οδηγήσει στη σημαντική ανάπτυξη της θεωρίας την τελευταία δεκαετία, όπου μια ευρεία ποικιλία αλγορίθμων και μεθόδων έχουν προταθεί. Ωστόσο, παρά την πρωτοφανή τους επιτυχία οι CF μέθοδοι παρουσιάζουν κάποιους σημαντικούς περιορισμούς συμπεριλαμβανομένης της επεκτασιμότητας και της αραιότητας των δεδομένων. Τα προβλήματα αυτά επιδρούν αρνητικά στην ποιότητα των παραγόμενων συστάσεων και διακυβεύουν την εφαρμοσιμότητα πολλών CF αλγορίθμων σε ρεαλιστικά σενάρια.
Χτίζοντας πάνω στη διαίσθηση πίσω από τον αλγόριθμο NCDawareRank - μίας γενικής μεθόδου υπολογισμού διανυσμάτων κατάταξης ιεραρχικά δομημένων γράφων - και της σχετικής με αυτόν έννοιας της NCD εγγύτητας, προβαίνουμε σε μία μοντελοποίηση του συστήματος με τρόπο που φωτίζει τα ενδημικά του χαρακτηριστικά και προτείνουμε έναν νέο αλγοριθμικό πλαίσιο συστάσεων, τον Αλγόριθμο 1. Στο επίκεντρο της προσέγγισής μας είναι η προσπάθεια να συνδυάσουμε τις άμεσες με τις NCD, ``γειτονιές'' των αντικειμένων ώστε να πετύχουμε μεγαλύτερης ακρίβειας χαρακτηρισμό των πραγματικών συσχετισμών μεταξύ των στοιχείων του χώρου αντικειμένων, με σκοπό την βελτίωση της ποιότητας των συστάσεων αλλά και την αντιμετώπιση της εγγενούς αραιότητας και των προβλημάτων που αυτή συνεπάγεται.
Για να αξιολογήσουμε την απόδοση της μεθόδου μας υλοποιούμε και εφαρμόζουμε τον Αλγόριθμο 1 στο κλασικό movie recommendation πρόβλημα και παραθέτουμε μια σειρά από πειράματα χρησιμοποιώντας τo MovieLens Dataset. Τα πειράματά μας δείχνουν πως ο Αλγόριθμος 1 με την εκμετάλλευση της ιδέας της NCD εγγύτητας καταφέρνει να πετύχει λίστες συστάσεων υψηλότερης ποιότητας σε σύγκριση με τις άλλες state-of-the-art μεθόδους που έχουν προταθεί στη βιβλιογραφία, σε ευρέως χρησιμοποιούμενες μετρικές (micro- και macro-DOA), αποδεικνύοντας την ίδια στιγμή πως είναι λιγότερο επιρρεπής στα προβλήματα που σχετίζονται με την αραιότητα και έχοντας παράλληλα ανταγωνιστικό προφίλ πολυπλοκότητας και απαιτήσεις αποθήκευσης.The purpose of this master's thesis is to study and develop a new algorithmic framework for collaborative filtering (CF) to generate recommendations. The method we propose is based on the exploitation of the hierarchical structure of the item space and intuitively ``stands'' on the property of Near Complete Decomposability (NCD) which is inherent in the structure of the majority of hierarchical systems. Collaborative Filtering is one of the most successful families of recommendations methods. The great impact of CF on Web applications, and its wide deployment in important commercial environments, have led to the significant development of the theory, with a wide variety of algorithms and methods being proposed. However, despite their unprecedented success, CF methods present some important limitations including scalability and data sparsity. These problems have a negative impact of the quality of the recommendations and jeopardize the applicability of many CF algorithms in realistic scenarios. Building on the intuition behind the NCDawareRank algorithm and its related concept of NCD proximity, we model our system in a way that illuminates its endemic characteristics and we propose a new algorithmic framework for recommendations, called Algorithm 1. We focus on combining the direct with the NCD `` neighborhoods'' of items to achieve better characterization of the inter-item relations, in order to improve the quality of recommendations and alleviate sparsity related problems. To evaluate the merits of our method, we implement and apply Algorithm 1 in the classic movie recommendation problem, running a number of experiments on the standard MovieLens dataset. Our experiments show that Algorithm 1 manages to create recommendation lists with higher quality compared with other state-of-the-art methods proposed in the literature, in widely used metrics (micro- and macro- DOA), demonstrating at the same time that it is less prone to low density related problems being at the same time very efficient in both complexity and storage requirements.
38
Kuo, Nai-Hao, and 郭乃豪. "Using Virus Optimization Algorithm on Collaborative Filtering Recommender System." Thesis, 2018. http://ndltd.ncl.edu.tw/handle/r4mqbx.
Full textAbstract:
碩士元智大學
工業工程與管理學系
106
Collaborative filtering, also known as CF, used the active data to analyze the preference of the user, and tried to make a recommendation. There are three kinds of mechanisms to create CF recommender systems: user-based, item-based, and model-based. The user-based and item-based CF systems use the rating data or basic data of users to analyze the correlation among users and provide the recommendation in real time. Yet, the model-based CF uses the historical data to train the model and offers the recommendation after the modeling is finished. In the past year, researchers are trying to create new methods to establish the recommender system or to improve the performance of the existing system. Some researches employed metaheuristic algorithms such as genetic algorithm or artificial immune system for the recommender system. Take the advantage of the complex mechanism, these metaheuristics are usually able to provide better prediction precision and higher efficiency. In this thesis, a recently developed metaheuristic algorithm, called virus optimization algorithm, is proposed to combine with collaborative filtering concept for the application of the recommender system. The popular movie recommendation database, MovieLens, is used to validate the performance of the proposed algorithm. A detailed design of experiments are implemented to find the best parameters of the algorithms and database. The best performance is then compared with several methods in the literature. The results show that the proposed VOA_RS outperforms most of algorithms in the literature and performs competitively to the rest. This study has successfully shown the merit of the VOA on the recommender system and its related applications.
39
Cheng, Hung-Lien, and 程閎廉. "A Hybrid Collaborative Filtering Recommender System Based on Clustering Algorithm." Thesis, 2010. http://ndltd.ncl.edu.tw/handle/52770657142827560926.
Full textAbstract:
碩士國立中興大學
資訊科學與工程學系所
98
Collaborative recommender is one of the most popular recommendation techniques. Traditional collaborative filtering approach mainly employs a matrix of user’s ratings on items to calculate the similarity between users. If the features of users or items are provided in the data set in addition to the rating data, then those features can be used to improve the quality of recommendations. In this thesis, we proposed a hybrid recommender system based on clustering and collaborative filtering techniques. In the proposed system, items are clustered based on item features and user-item rating matrix. Similarly, users are clustered based on the user’s preferred categories of items and user-item rating matrix. Then a hybrid method that combines content-based and collaborative filtering is proposed to predict the rating of an item for a given user. The experimental results show that the proposed method has higher accuracy in terms of mean absolute error than that of User-based collaborative filtering approach, Item-based filtering approach, Clustering Items for Collaborative Filtering (CICF), and the User Profile Clustering (UPC) method. Especially, when the dataset is sparse, the accuracy of the proposed method is better and more stable than the other methods.
40
Cunha, Tiago Daniel Sá. "Recommending Recommender Systems: tackling the Collaborative Filtering algorithm selection problem." Doctoral thesis, 2019. https://hdl.handle.net/10216/125150.
Full text
41
Cunha, Tiago Daniel Sá. "Recommending Recommender Systems: tackling the Collaborative Filtering algorithm selection problem." Tese, 2019. https://hdl.handle.net/10216/125150.
Full text
42
RATHI, ISHAN. "A COLLABORATIVE FILTERING-BASED RECOMMENDER SYSTEM ALLEVIATING COLD START PROBLEM." Thesis, 2019. http://dspace.dtu.ac.in:8080/jspui/handle/repository/16694.
Full textAbstract:
Consumers currently have a surplus of items available to purchase via online stores. Surplus of goods enables users to have huge variety but it often leads to inconvenience for users. Consumers have to spend a lot of time going through items to find goods of their preference. To automate the process of sharing relevant suggestions, recommender systems are used.
Recommender systems are making their presence felt in a number of domains, be it for ecommerce or education, social networking etc. With huge growth in number of consumers and items in recent years, recommender systems face some key challenges. These are: producing high quality recommendations and performing many recommendations per second for millions of consumers and items. New recommender system technologies are needed to scale themselves for new items as well as in new user in the system in order to get high quality recommendations.
In this thesis, we focus on collaborative approach-based recommender systems to solve the issue of cold start problem. We have compared multiple algorithms which aim to solve cold start problem and proposed a new hybrid algorithm. This new algorithm is implemented on Movie-Lens 1Million Dataset.
43
Zhang, Rui-zhe, and 張睿哲. "A Group Recommender System based on Neural Network Collaborative Filtering Algorithm." Thesis, 2011. http://ndltd.ncl.edu.tw/handle/ar54r9.
Full textAbstract:
碩士國立臺灣科技大學
電機工程系
99
In the research field of Recommender System, the personalized recommender system has always been regarded as the main trend. The main operation is to record the pattern of personal behavior in the system and then the system will recommend users which program to use according to users’ preferences. However, consumer behaviors and recreational activities are not both formed by one single individual, many of which will be made by groups. For example, when relatives and friends get together to go to see movies, go travelling or having meals, the single user recommender system can’t achieve the application targets for above situations. In the past, the group recommender system intended to combine users’ preference on the same aspect to achieve different variety of measurement. However, this approach ignores individual member’s characteristics and the pattern of how each member interacts with others; such that this sort of measurement can’t truly reflect real interest on the same issue. The main target of our research is to develop a group recommender system based on neural network training algorithms. We proposed to train and get weights in the neural network and to simulate phenomena made by the interaction among groups by measuring the same issue made by each individual or the whole group. Finally, we evaluate our approach experimentally and compare it in different parameter of network. The experimental result shows that we can achieve the function of being user-friendly by algorithm in the group recommender system that can’t be achieve by algorithm in the personalized recommender system.
44
Chen, Chih-Ta, and 陳志達. "Applying Anytime Algorithm and Progressive Reasoning to Collaborative Filtering System Design." Thesis, 2005. http://ndltd.ncl.edu.tw/handle/66173143503423182794.
Full textAbstract:
碩士國立屏東商業技術學院
資訊管理系
93
Problems of information overload become much critical since internet being an essential path for information absorption. Therefore, different kinds of information recommendation systems are developed to solve the problems. This study presents a recommendation system design of progressive reasoning using collaborative filtering method. Progressive reasoning system can decrease the doubt of data privacy, however, its computation length cause inconvenience for further application. The study aims to employ collaborative filtering to recommend information of likely preference based on the experience of near neighbor. Then use the progressive reasoning to locate the real preference by users themselves. The searching of near neighbor is conducted by computing the similarity between new user and existing users. In the recommendation system anytime algorithms provide the capability to trade deliberation time for quality of results. A prototype also presents in the study where the data sample are collected from newsletter of a website.
45
Tsai, Ming-Huang, and 蔡明晃. "A matrix factorization-based estimation of distribution algorithm for collaborative filtering recommender systems." Thesis, 2014. http://ndltd.ncl.edu.tw/handle/72101814750921203603.
Full textAbstract:
碩士國立高雄第一科技大學
電腦與通訊工程研究所
102
Collaborative filtering is one of the crucial components of recommender systems.By analyzing relationships between customers and products, this component is able to identify new user-item associations, to predict user preferences, and to generate product recommendations. With a large number of users, however, the use of collaborative filtering for producing recommendations is computationally expensive. This study therefore proposes a matrix factorization-based estimation of distribution algorithm for collaborative filtering recommender systems.Performance of the proposed algorithm is evaluated by comparing it against other competitive approaches. Experimental results using the MovieLens dataset demonstrate the practicality of the proposed algorithm.
46
Chu, Po-Wei, and 朱栢葦. "Implementation and Performance Evaluation of a Collaborative Filtering Recommendation Algorithm on the Cloud Using Hadoop." Thesis, 2015. http://ndltd.ncl.edu.tw/handle/83980248134088009171.
Full textAbstract:
碩士東海大學
資訊工程學系
103
With the rapid development of the cloud computing technology and the Internet, how to help users quickly gain valuable insights from the Big Data has gradually become the focus and challenge. From analysis of user log information on the Internet, preferences of the user are revealed, information or commodities most matched with user interests are recommended. However, with the fast development of the Internet, the amount of user log files explodes, causing bottlenecks in performance and storage spaces for recommendation systems. To solve this problem, in this thesis we implement a recommendation system architecture based on the Hadoop distributed computing framework, in which the Hadoop distributed file storage (HDFS), distributed computing frameworks MapReduce, and Mahout machine-learning algorithm based database are adopted. We also explore performance evaluation of the collaborative filtering recommendation algorithm on the cloud cluster and standalone personal computer.
47
Chou, Yi-Ching, and 周怡青. "Applying Deep Graphical Collaborative Filtering Algorithm to Analyze Trading Behavior Sequence for Mutual Fund Recommendation." Thesis, 2019. http://ndltd.ncl.edu.tw/handle/79585e.
Full textAbstract:
碩士國立交通大學
資訊管理研究所
107
In recent years, many industries start applying machine learning and deep learning technology to solve complex problems with the development of artificial intelligence. Financial services also tend to be more intelligent and personalized, while precision marketing is usually the ultimate goal of personalized recommendation. However, with the great number of customers and products but the sparse transaction records in the real-world cases, it comes up with data sparsity, model scalability, cold start problems which makes accurate personalized recommendations be a tough issue. Moreover, when the number of customers and products increases massively with time, the scale scalability of the recommender model will affect its training and recommendation efficiency. In this paper, we propose a Graphical Deep Collaborative Filtering (GraphDCF) algorithm to analyze the trading behavior sequence of customers and provide personalized mutual fund recommendations. The graph-structured network is constructed based on connecting customer nodes with similar purchasing and redeeming trading order in the sequential view. Next, generating the corresponding embedding vector for each customer node based on similar transaction behaviors by operating the aggregate function. Finally, we combined the proposed Deep Embedded Collaborative Filtering (DECF) framework to predict the willingness of purchasing specific mutual funds for each customer and provide personalized recommendations. The experimental results show that our GraphDCF algorithm can improve the recommendation performance up to 2.3% effectively and the training efficiency of the recommendation model with maintaining the precision measurement.
48
Tsai, Chang-You, and 蔡長祐. "Hybrid Collaborative Filtering Recommendation System Using Genetic Algorithm-based K-means Clustering : The Case of Airlines." Thesis, 2018. http://ndltd.ncl.edu.tw/handle/t2zz6w.
Full textAbstract:
碩士國立臺北科技大學
工業工程與管理系
106
According to statistics, the airline industry has accumulated a new record 4.1 billion passengers on regular flights, and global passenger traffic has grown faster than the average of the past decade. The airlines through differentiated services and fares to win customers in different markets, but for consumers, they will hope the airlines have better services and feedback that they take, so that they always spend a lot of time to look for then delay the journey. Therefore, this study will develop a hybrid collaborative filtering recommendation system using GA K-means clustering. First, this study will collect airlines rating data and make an airline service performance evaluation table based on the five criteria of price, cabin space, passenger services, inflight entertainment, and meals. Then the genetic algorithm-based K-means clustering was grouped by similar airlines. Collaborative filtering based on the ratio of the group that has been grouped and the users preference based on the airlines preferences, to find out the cluster to which the user belongs. After confirming the cluster to which the user belongs, the eligible airline is extracted through what the user enters. Then, using ratio-based collaborative filtering, find the airline that is similar to the user’s preference for the airline and that match the location where the user want to go that can recommend for the user which airline is fitting. Finally, this study will use the data from a rating website and a location where to go that illustrate and verify the feasibility and practical value of the study.