Global ETD Search

1	MetaStackVis: Visually-Assisted Performance Evaluation of Metamodels in Stacking Ensemble Learning Ploshchik, Ilya January 2023 (has links) Stacking, also known as stacked generalization, is a method of ensemble learning where multiple base models are trained on the same dataset, and their predictions are used as input for one or more metamodels in an extra layer. This technique can lead to improved performance compared to single layer ensembles, but often requires a time-consuming trial-and-error process. Therefore, the previously developed Visual Analytics system, StackGenVis, was designed to help users select the set of the most effective and diverse models and measure their predictive performance. However, StackGenVis was developed with only one metamodel: Logistic Regression. The focus of this Bachelor's thesis is to examine how alternative metamodels affect the performance of stacked ensembles through the use of a visualization tool called MetaStackVis. Our interactive tool facilitates visual examination of individual metamodels and metamodels' pairs based on their predictive probabilities (or confidence), various supported validation metrics, and their accuracy in predicting specific problematic data instances. The efficiency and effectiveness of MetaStackVis are demonstrated with an example based on a real healthcare dataset. The tool has also been evaluated through semi-structured interview sessions with Machine Learning and Visual Analytics experts. In addition to this thesis, we have written a short research paper explaining the design and implementation of MetaStackVis. However, this thesis provides further insights into the topic explored in the paper by offering additional findings and in-depth analysis. Thus, it can be considered a supplementary source of information for readers who are interested in diving deeper into the subject. Visualization interaction metamodels validation metrics predicted probabilities stacking stacked generalization ensemble learning machine learning Computer Sciences Datavetenskap (datalogi)
2	A Logistic Regression Analysis of Utah Colleges Exit Poll Response Rates Using SAS Software Stevenson, Clint W. 27 October 2006 (has links) (PDF) In this study I examine voter response at an interview level using a dataset of 7562 voter contacts (including responses and nonresponses) in the 2004 Utah Colleges Exit Poll. In 2004, 4908 of the 7562 voters approached responded to the exit poll for an overall response rate of 65 percent. Logistic regression is used to estimate factors that contribute to a success or failure of each interview attempt. This logistic regression model uses interviewer characteristics, voter characteristics (both respondents and nonrespondents), and exogenous factors as independent variables. Voter characteristics such as race, gender, and age are strongly associated with response. An interviewer's prior retail sales experience is associated with whether a voter will decide to respond to a questionnaire or not. The only exogenous factor that is associated with voter response is whether the interview occurred in the morning or afternoon. Exit Poll sampling weighting questionnaires interviewers voters polling places precincts backward selection SAS proc logistic surveylogistic model selection Utah Colleges Exit Poll UCEP SUDAAN SQL data management weight adjustments predicted probabilities logistic regression holdout research validation sample design data architecture adjustment exogenous statistics political science voters nonresponse response rate retail sales age race gender time of day Statistics and Probability

Search results

MetaStackVis: Visually-Assisted Performance Evaluation of Metamodels in Stacking Ensemble Learning

A Logistic Regression Analysis of Utah Colleges Exit Poll Response Rates Using SAS Software