Добірка наукової літератури з теми "Holt-Winter method"

Exponential smoothing is one of the short term forecasting methods. The selection of the forecasting method can be done by considering the type of data pattern, such as the Double Exponential Smoothing (DES) Holt method which can be used on trend patterned data and the Triple Exponential Smoothing (TES) Holt-Winter method which can be used on trend and seasonal patterned data. The main problem in using the Holt DES and Holt-Winter TES methods is the parameter selection which is usually done by trial and error, but this method takes a long time so that in this research a more efficient method is used to obtain optimal parameters, namely the golden section method. The purpose of this research was to forecast and obtain the best method for forecasting the export value of East Borneo Province. The results showed that the forecasted of export values used the Holt DES, the additive Holt-Winter TES, and the multiplicative Holt-Winter TES with golden section optimization method had a MAPE of less than 10%, which means that the forecast used these methods were very good. The best method to predict the export value of East Borneo Province was the additive Holt-Winter TES with golden section optimization method.

Rice, as a staple food, plays a crucial role in global food security. Accurate forecasting of rice prices is essential for policymakers, farmers, and consumers alike. This article explores the application of the Holt-Winter exponential smoothing model to predict rice prices. Holt-Winter method is chosen for its ability to capture both trend and seasonality in time series data, which are prominent features in agricultural commodity prices such as rice. The study analyzes historical price data, identifies trends, seasonality, and incorporates smoothing parameters in additive and multiplicative methods. Results indicate that additive method of Holt-Winter exponential smoothing provides a better performance. This research contributes valuable insights to the field of agricultural economics and informs strategies for managing food supply chains and market stability.

This paper mainly forecasts the daily closing price of stock markets. We propose a two-stage technique that combines the empirical mode decomposition (EMD) with nonparametric methods of local linear quantile (LLQ). We use the proposed technique, EMD-LLQ, to forecast two stock index time series. Detailed experiments are implemented for the proposed method, in which EMD-LPQ, EMD, and Holt-Winter methods are compared. The proposed EMD-LPQ model is determined to be superior to the EMD and Holt-Winter methods in predicting the stock closing prices.

UD. Fajar Jaya is a trading business unit engaged in the supply of souvenirs. But in the management of the business there are some problems of which are UD. Fajar Jaya can not predict how the optimal number of souvenirs that must be provided to customers on every item souvenirs are sold. This causes the service to consumers less than the maximum, especially at certain moments sales of souvenirs (example: glass souvenirs) jumped dramatically from the number of average sales. To overcome the above, the authors propose to forecast the level of sales of souvenirs using Holt and Winter methods that exist in the development of Exponential Smoothing (ES) method. From the application of the two methods, then will make comparison of effectiveness of method which measured through actual data accuracy and forecasting result by knowing forecast error level. From the research results obtained forecasting results for Holt Double Exponential Smoothing method in July of 2017 is amounted to 599 items that may be sold with MAD forecasting error rate of 10.54 and MAPE of 3.70%. As for forecasting using Winter Exponential Smoothing method in July of 2017 is 549.6 items that may be sold with MAD 0.02 and MAPE error rate of 2.55%. The conclusion that can be drawn from the research that has been done on sales data souvenirs on UD. Fajar Jaya is that the Winter Exponential Smoothing method is more suitable to be applied in case study of souvenir sales in UD. Fajar Jaya is more than Holt Double Exponential Smoothing method.

Ujung Genteng Sukabumi Beach is one of the tourism destinations in Sukabumi Regency, West Java. Forecasting tourist arrivals is a very important factor for tourist destination policies and contributes to the regional economy and the surrounding community. The purpose of this study is to predict the number of tourists who come to Ujung Genteng Beach, Sukabumi. The method used is the Holt-Winter approach exponential smoothing. The Holt-Winter method is used for data that is not stationary, has both trend and seasonal elements. The Holt-Winters method has two models, namely the Additive model and the Multiplicative model. The data used is visitor data in January 2017 - February 2020, the results of the analysis show that the prediction of the number of visitors to Ujung Genteng beach in March 2020 from the additive model is 300 people with a MAPE value of 85.48% and an MSE value of 31230672.68 and a prediction of the number of beach visitors. Ujung Genteng in March 2020 from a multiplicative model of 740 people, with MAPE and MSE values obtained were 86.34% and 27754873.34.

The tourism industry in Malaysia has been growing significantly over the years. Tourism has been one of the major donors to Malaysia’s economy. Based on the report from the Department of Statistics, a total of domestic visitors in Malaysia were recorded at about 221.3 million in 2018 with an increase of 7.7% alongside a higher record in visitor arrivals and tourism expenditure. This study aims to make a comparison between two methods, which are Fuzzy Time Series and Holt-Winter in forecasting the number of tourist arrival in Langkawi based on the monthly tourist arrival data from January 2015 to December 2019. Both models were generated using Microsoft Excel in obtaining the forecast value. The Mean Square Error (MSE) has been calculated in this study to get the best model by looking at the lowest value. The result found that Holt-Winter has the lowest value that is 713524285 compared to the Fuzzy Time Series with a value of 2625517469. Thus, the Holt-Winter model is the best method and has been used to forecast the tourist arrival for the next 2 years. The forecast value for the years 2020 and 2021 are displayed by month.

The competition level between companies on executing product marketing is rapidly increasing, so the companies have to understand the importance of correlation between external environments of company with consumer’s needs. One of the efforts that can be done is by utilizing data warehouse and the application of infrastructure in information and technology field. This research combined Association Rules method to extracting pattern and finding every possibility that potential to increase sales and Holt-Winter Multiplicative method to estimate the alteration of trend on the seasonal data. After passed through data processing process by using RapidMiner tools, information that consists of correlation pattern between rule that describe the comparison of product and the sales working area and season that affects the product sale. The pattern used by company to know which product is often purchased by customer. Besides that, this research produces changing trend data of PT ABC’s product that generated by result of previous data comparison with forecast data. Based on value of error rate Mean Absolute Percentage Error (MAPE) in estimating forecast result on the PT ABC’s sales transaction data during 3 years, it shows good level of accuracy. Result of data test, by considering rule that formed and forecast result so the company can control and manage product in order to avoid incorrect sales. This thing will effect on repression of operational cost and PT ABC can identify available opportunities to increase sale of agricultural medicine.

Consumer demand is an important factor in any business, especially in the food retail industry whose products are perishable and have a short life cycle. The daily demand for a food product is affected by external factors, such as seasonality, price reduction and holidays. To satisfy the stochastic demand, product characteristics vary with customer are required to be timely updated based on market dynamicity. According to previous research, to choose suitable forecasting model is the main concern of enterprises on demand management issue. Proper demand forecasting provides organization with valuable information regarding their prospective in their current market, allowing to make appropriated production portfolio. By applying ARIMA and Holt-Winter, this paper aims to forecast the canned fruit demand at a specific company to help them eliminate waste of lean related to production and distribution. Results are evaluated according to forecasting errors (MAD, MSE, MAPE). By comparing the aforementioned methods, it can be concluded that ARIMA outperforms Holt-Winter related to prediction accuracy.

Penelitian ini bertujuan untuk mengembangkan produk Forecasting System Multi-Model (FSM) guna menentukan metode terbaik dalam sistem peramalan (forecast) dengan mengkonstruksi beberapa metode dalam bentuk Graphical User Interface (GUI) Matlab dengan menghitung semua indikator tingkat akurasi guna menemukan model matematika terbaik dari data time series pada periode tertentu. Pada tahap simulasi, tim peneliti menggunakan data Indeks Pembangunan Manusia (IPM) Provinsi Nusa Tenggara Barat (NTB) tahun 2010-2017 guna memprediksi IPM NTB tahun 2018. Adapun metode yang diuji adalah Moving Average (SMA, WMA dan EMA), Exponential Smoothing Method (SES, Brown, Holt, dan Winter), Naive Method, Interpolation Method (Newton Gregory), dan Artificial Neural Network (Back Propagation). Kemudian model dievaluasi untuk melihat tingkat akurasi masing-masing metode berdasarkan nilai MAD, MSE, dan MAPE. Berdasarkan hasil simulasi data dari 10 metode yang diuji diketahui bahwa metode Holt paling akurat dengan hasil prediksi tahun 2018 sebesar 67,45 dengan MAD, MSE, dan MAPE berturut-turut sebesar 0,22654; 0,075955 dan 0,34829. The purpose of this research is to develop a product was called Forecasting System Multi-Model (FSM) to determine the best method in the forecasting system by constructing several methods in the form of Graphical User Interface (GUI) Matlab. It was done by all indicator accuration to find the best mathematical model of time series data in a certain period. In the simulation phase, this research used the Human Development Index (HDI) data of West Nusa Tenggara (NTB) Province in 2010 - 2017 to predict the HDI data of NTB in 2018. The methods tested were Moving Average (SMA, WMA and EMA), Exponential Smoothing Method (SES, Brown, Holt, and Winter), Naive Method, Interpolation Method (Newton Gregory), and Artificial Neural Network (Back Propagation). Then the models/methods were evaluated to see the level of accuracy of each method based on the value of MAD, MSE, and MAPE. Based on data simulation result from 10 tested method known that Holt method is most accurate with prediction result of 2018 equal to 67,45 with MAD, MSE, and MAPE respectively equal to 0.22654, 0.075955 and 0.34829.

Weather conditions in the city of Mataram tend to be erratic and difficult to predict, such as the condition of rainfall data in 2018 which changes over a certain period of time so that the weather is difficult to predict accurately. In this study, we propose the Exponential Smoothing Holt-Winter method to forecast rainfall in the city of Mataram, so that it can be a decision support for various interested sectors. This method has been tested using secondary data from the Mataram City Central Bureau of Statistics for the period January 2014 to 2018 and evaluated using Mean Absolute Deviation (MAD), Mean Squared Error (MSE) and Mean Absolute Percentage Error (MAPE). The results of this study indicate that using the Exponential Smoothing Holt-Winter method yields better results, each of which is MAPE 142.3, MAD 95.6 and MSD value 24988.7 and the data smoothing value is obtained for the smallest combination value of α 0.2, β 0.1, and γ 0.1. It can be concluded that the proposed method can provide better information and can be used to predict rainfall in Mataram City for the next 12 periods.

I denna rapport presenteras ett tillvägagångssätt för att hitta och datera brytpunkter i tidsserier. En brytpunkt definieras av det datum då det skett en stor nivåförändring i tidsserien. Det presenteras även en strategi för att skatta effekten av daterade brytpunkter. Genom att analysera tidsserier över AFA Försäkrings ärendeinflöde visar det sig att brytpunkter i tidsserien sammanfaller med exogena händelser som kan ha orsakat dessa brytpunkter, till exempel villkors- eller lagförändringar inom försäkringsbranschen. Rapporten visar att det genom ett metodiskt angreppssätt går att skatta effekten av en exogen händelse. Dessa skattade effekter kan användas vid framtida prognoser då en liknande förändring förväntas inträffa. Dessutom skapas prognoser över ärendeinflödet två år framåt med olika tidsseriemodeller.

Le rôle clé que l’IA pourrait jouer dans l’amélioration des activités commerciales est connu depuis longtemps, mais le processus de pénétration de cette nouvelle technologie a rencontré certains freins au sein des entreprises, en particulier, les coûts de mise œuvre. En moyenne, 2.8 ans sont nécessaires depuis la sélection du fournisseur jusqu’au déploiement complet d’une nouvelle solution. Trois points fondamentaux doivent être pris en compte lors du développement d’un nouveau modèle. Le désalignement des attentes, le besoin de compréhension et d’explications et les problèmes de performance et de fiabilité. Dans le cas de modèles traitant des données de la supply chain, cinq questions spécifiques viennent s’ajouter aux précédentes : - La gestion des incertitudes. Les décideurs cherchent un moyen de minimiser le risque associé à chaque décision qu’ils doivent prendre en présence d’incertitude. Obtenir une prévision exacte est un rêve ; obtenir une prévision assez précise et en calculer les limites est réaliste et judicieux. - Le traitement des données entières et positives. La plupart des articles ne peuvent pas être vendus en sous-unités. Cet aspect simple de la vente se traduit par une contrainte qui doit être satisfaite : le résultat doit être un entier positif. - L’observabilité. La demande du client ne peut pas être mesurée directement, seules les ventes peuvent être enregistrées et servir de proxy. - La rareté et la parcimonie. Les ventes sont une quantité discontinue. En enregistrant les ventes par jour, une année entière est condensée en seulement 365 points. De plus, une grande partie d’entre elles sera à zéro. - L’optimisation juste-à-temps. La prévision est une fonction clé, mais elle n’est qu’un élément d’une chaîne de traitements soutenant la prise de décision. Le temps est une ressource précieuse qui ne peut pas être consacrée entièrement à une seule fonction. Le processus de décision et les adaptations associées doivent donc être effectuées dans un temps limité et d’une manière suffisamment flexible pour pouvoir être interrompu et relancé en cas de besoin afin d’incorporer des événements inattendus ou des ajustements nécessaires. Cette thèse s’insère dans ce contexte et est le résultat du travail effectué au cœur de Lokad. La recherche doctorale a été financée par Lokad en collaboration avec l’ANRT dans le cadre d’un contrat CIFRE. Le travail proposé a l’ambition d’être un bon compromis entre les nouvelles technologies et les attentes des entreprises, en abordant les divers aspects précédemment présentés. Nous avons commencé à effectuer des prévisions en utilisant la famille des lissages exponentiels, qui sont faciles à mettre en œuvre et extrêmement rapides à exécuter. Largement utilisés dans l’industrie, elles ont déjà gagné la confiance des utilisateurs. De plus, elles sont faciles à comprendre et à expliquer à un public non averti. En exploitant des techniques plus avancées relevant du domaine de l’IA, certaines des limites des modèles utilisés peuvent être surmontées. L’apprentissage par transfert s’est avéré être une approche pertinente pour extrapoler des informations utiles dans le cas où le nombre de données disponibles était très limité. Nous avons proposé d’utiliser un modèle associé à une loi de Poisson, une binomiale négative qui correspond mieux à la nature des phénomènes que nous cherchons à modéliser et à prévoir. Nous avons aussi proposé de traiter l’incertitude par des simulations de Monte Carlo. Un certain nombre de scénarios sont générés, échantillonnés et modélisés par dans une distribution. À partir de cette dernière, des intervalles de confiance de taille différentes et adaptés peuvent être déduits. Sur des données réelles de l’entreprise, nous avons comparé notre approche avec les méthodes de l’état de l’art comme DeepAR, DeepSSMs et N-Beats. Nous en avons déduit un nouveau modèle conçu à partir de la méthode Holt-Winter [...]
The key role that AI could play in improving business operations has been known for a long time, but the penetration process of this new technology has encountered certain obstacles within companies, in particular, implementation costs. On average, it takes 2.8 years from supplier selection to full deployment of a new solution. There are three fundamental points to consider when developing a new model. Misalignment of expectations, the need for understanding and explanation, and performance and reliability issues. In the case of models dealing with supply chain data, there are five additionally specific issues: - Managing uncertainty. Precision is not everything. Decision-makers are looking for a way to minimise the risk associated with each decision they have to make in the presence of uncertainty. Obtaining an exact forecast is a advantageous; obtaining a fairly accurate forecast and calculating its limits is realistic and appropriate. - Handling integer and positive data. Most items sold in retail cannot be sold in subunits. This simple aspect of selling, results in a constraint that must be satisfied by the result of any given method or model: the result must be a positive integer. - Observability. Customer demand cannot be measured directly, only sales can be recorded and used as a proxy. - Scarcity and parsimony. Sales are a discontinuous quantity. By recording sales by day, an entire year is condensed into just 365 points. What’s more, a large proportion of them will be zero. - Just-in-time optimisation. Forecasting is a key function, but it is only one element in a chain of processes supporting decision-making. Time is a precious resource that cannot be devoted entirely to a single function. The decision-making process and associated adaptations must therefore be carried out within a limited time frame, and in a sufficiently flexible manner to be able to be interrupted and restarted if necessary in order to incorporate unexpected events or necessary adjustments. This thesis fits into this context and is the result of the work carried out at the heart of Lokad, a Paris-based software company aiming to bridge the gap between technology and the supply chain. The doctoral research was funded by Lokad in collaborationwith the ANRT under a CIFRE contract. The proposed work aims to be a good compromise between new technologies and business expectations, addressing the various aspects presented above. We have started forecasting using the exponential smoothing family which are easy to implement and extremely fast to run. As they are widely used in the industry, they have already won the confidence of users. What’s more, they are easy to understand and explain to an unlettered audience. By exploiting more advanced AI techniques, some of the limitations of the models used can be overcome. Cross-learning proved to be a relevant approach for extrapolating useful information when the number of available data was very limited. Since the common Gaussian assumption is not suitable for discrete sales data, we proposed using a model associatedwith either a Poisson distribution or a Negative Binomial one, which better corresponds to the nature of the phenomena we are seeking to model and predict. We also proposed using Monte Carlo simulations to deal with uncertainty. A number of scenarios are generated, sampled and modelled using a distribution. From this distribution, confidence intervals of different and adapted sizes can be deduced. Using real company data, we compared our approach with state-of-the-art methods such as DeepAR model, DeepSSMs and N-Beats. We deduced a new model based on the Holt-Winter method. These models were implemented in Lokad’s work flow