Auswahl der wissenschaftlichen Literatur zum Thema „Ridge leverage scores“

ImportanceStudents who ride older school buses are often exposed to high levels of exhaust during their commutes, which may adversely affect health and school attendance. As a result, the US Environmental Protection Agency (EPA) has awarded millions of dollars to school districts to replace older, highly polluting school buses with newer, cleaner buses.ObjectiveTo leverage the EPA’s randomized allocation of funding under the 2012-2016 School Bus Rebate Programs to estimate the association between replacing old, highly polluting buses and changes in district-average standardized test scores.Design, Setting, and ParticipantsThis study examined changes in reading and language arts (RLA) and math test scores among US school district applicants to the EPA’s 2012-2016 national School Bus Rebate Programs 1 year before and after each lottery by selection status. Data analysis was conducted from January 15 to July 30, 2023.ExposureSelection to receive EPA funding to replace older school buses with newer, cleaner alternatives.Main Outcomes and MeasuresSchool district changes in RLA and math test scores among students in grades 3 through 8 before and after the EPA funding lotteries by selection status were measured using an intention-to-treat approach.ResultsThis study included 1941 school district applicants to the 2012-2106 EPA School Bus Rebate Programs. These districts had a mean (SD) of 14.6 (33.7) schools per district, 8755 (23 776) students per district, and 41.3% (20.2%) of students with free lunch eligibility. Among the applicants, 209 districts (11%) were selected for the clean bus funding. District-average student test scores did not improve among selected districts overall. In secondary analyses, however, districts replacing the oldest, highest polluting buses (ie, pre-1990) experienced significantly greater improvements in district-average test scores in the year after the lottery for RLA and math (SD improvement in test scores, 0.062 [95% CI, 0.050-0.074] and 0.025 [95% CI, 0.011-0.039], respectively) compared with districts without replacements.Conclusions and RelevanceIn this study, the EPA funding was not associated with student test scores overall, but in secondary analyses, the replacement of the oldest school buses was associated with improved educational performance. These findings support prioritizing clean bus replacement of the oldest buses as an actionable way for improving students’ educational performance.

This research aims to comprehensively investigate the dynamics of work-life balance and to develop predictive models using machine learning techniques to assess and predict the factors influencing work-life equilibrium. The study leverages a dataset containing 15,973 responses obtained from the global work-life survey conducted by Authentic-Happiness.com. The survey comprises 23 questions, providing a multifaceted view of how individuals manage their personal and professional lives. Initial Exploratory Data Analysis (EDA) uncovers five key dimensions: "Healthy Body," "Healthy Mind," "Expertise," "Connection," and "Meaning." These dimensions are explored to gain insights into their significance in relation to work-life balance. Subsequently, an extensive set of machine learning regression models, including Linear Regression, Decision Tree Regression, Random Forest Regression, Gradient Boosting Regressor, XGBoost, LightGBM, CatBoost, Support Vector Machine, K Nearest Neighbors, K-Means Regression, Ridge and Lasso Regression, Principal Component Analysis, RANSAC, Quartile Regression, GAM, Huber Regression, RBF Kernel Regression, and SGD Regression, are employed to predict work-life balance scores. Performance evaluation is based on metrics such as Mean Squared Error (MSE) and R-squared (R²). The research uncovers a holistic understanding of work-life balance and identifies significant predictors. The comparative analysis of machine learning models reveals their effectiveness in predicting work-life balance, highlighting the models that perform optimally. This research contributes valuable insights into the intricate factors that underlie work-life balance, offering a data-driven perspective that can inform personal choices, organizational strategies, and policy decisions. The application of machine learning techniques underscores the potential for addressing contemporary challenges associated with achieving a harmonious work-life equilibrium.

AbstractPrincipal Component Analysis (PCA) is a well-known linear dimension-reduction technique designed for Euclidean data. In a wide spectrum of applied fields, however, it is common to observe multivariate circular data (also known as toroidal data), rendering spurious the use of PCA on it due to the periodicity of its support. This paper introduces Toroidal Ridge PCA (TR-PCA), a novel construction of PCA for bivariate circular data that leverages the concept of density ridges as a flexible first principal component analog. Two reference bivariate circular distributions, the bivariate sine von Mises and the bivariate wrapped Cauchy, are employed as the parametric distributional basis of TR-PCA. Efficient algorithms are presented to compute density ridges for these two distribution models. A complete PCA methodology adapted to toroidal data (including scores, variance decomposition, and resolution of edge cases) is introduced and implemented in the companion R package . The usefulness of TR-PCA is showcased with a novel case study involving the analysis of ocean currents on the coast of Santa Barbara.

Les contributions de cette thèse se divisent en deux parties. Une première partie dédiée à l’accélération des méthodes à noyaux et une seconde à l'optimisation sous contrainte de parcimonie. Les méthodes à noyaux sont largement connues et utilisées en apprentissage automatique. Toutefois, la complexité de leur mise en œuvre est élevée et elles deviennent inutilisables lorsque le nombre de données est grand. Nous proposons dans un premier temps une approximation des Ridge Leverage Scores. Nous utilisons ensuite ces scores pour définir une distribution de probabilité pour le processus d'échantillonnage de la méthode de Nyström afin d’accélérer les méthodes à noyaux. Nous proposons dans un second temps un nouveau framework basé sur les noyaux, permettant de représenter et de comparer les distributions de probabilités discrètes. Nous exploitons ensuite le lien entre notre framework et la Maximum Mean Discrepancy pour proposer une approximation précise et peu coûteuse de cette dernière. La deuxième partie de cette thèse est consacrée à l’optimisation avec contrainte de parcimonie pour l’optimisation de signaux et l’élagage de forêts aléatoires. Tout d’abord, nous prouvons sous certaines conditions sur la cohérence du dictionnaire, les propriétés de reconstruction et de convergence de l’algorithme Frank-Wolfe. Ensuite, nous utilisons l'algorithme OMP pour réduire la taille de forêts aléatoires et ainsi réduire la taille nécessaire pour son stockage. La forêt élaguée est constituée d’un sous-ensemble d’arbres de la forêt initiale sélectionnés et pondérés par OMP de manière à minimiser son erreur empirique de prédiction
The contributions of this thesis are divided into two parts. The first part is dedicated to the acceleration of kernel methods and the second to optimization under sparsity constraints. Kernel methods are widely known and used in machine learning. However, the complexity of their implementation is high and they become unusable when the number of data is large. We first propose an approximation of Ridge leverage scores. We then use these scores to define a probability distribution for the sampling process of the Nyström method in order to speed up the kernel methods. We then propose a new kernel-based framework for representing and comparing discrete probability distributions. We then exploit the link between our framework and the maximum mean discrepancy to propose an accurate and fast approximation of the latter. The second part of this thesis is devoted to optimization with sparsity constraint for signal optimization and random forest pruning. First, we prove under certain conditions on the coherence of the dictionary, the reconstruction and convergence properties of the Frank-Wolfe algorithm. Then, we use the OMP algorithm to reduce the size of random forests and thus reduce the size needed for its storage. The pruned forest consists of a subset of trees from the initial forest selected and weighted by OMP in order to minimize its empirical prediction error

Understanding human actions has been an important area of computer vision based deep learning domain. Several landmark extraction frameworks like media pipe and open Pose are used to extract the landmark coordinates from the body. The proposed work leverages open-source body landmark extraction and then trains a deep learning model on custom dataset created. The proposed work classifies the human body actions into blank face, yawn, namaste, punch and kick actions. The dataset creation phase involved recording of actions corresponding to every class and flattening them into a data frame. The dataset was later trained on a machine learning pipeline with machine learning algorithms like logistic regression, ridge classifier, random forest, and gradient boosting classifier. The algorithm with best accuracy was taken for real time usage. The landmark extraction model i.e., Mediapipe was used both in creation of dataset and execution of model in real time. The deep learning model was evaluated and validated based on several evaluation metrics like accuracy, confusion matrix, confidence score and recall score. The work proposed computationally efficient way of detecting the actions performed by the subject on camera by leveraging deep learning methods and mediapipe perception model for landmark extraction.