Random Forest vs Logistic Regression in Predictive Analytics Applications

Date
2018-11
Authors
Palacat, Christi
Contributor
Advisor
Department
Instructor
Wen, Eric
Depositor
Speaker
Researcher
Consultant
Interviewer
Annotator
Journal Title
Journal ISSN
Volume Title
Publisher
University of Hawaii - West Oahu
Volume
Number/Issue
Starting Page
Ending Page
Alternative Title
Abstract
Are there significant differences in prediction accuracy between modern data-mining approaches and classical regression methods? Which approach is easier to explain to enrollment managers when estimating student outcomes of varying levels of complexity? Complexity in the data is typically associated with quality, quantity, and the interaction of predictor variables. To test for such differences, this study compares the classification accuracy of a random forest algorithm with binomial logistic regression for purposes of predicting student admission yield. Findings are translated into operationally meaningful indicators in the context of enhanced institutional research on yield prediction and enrollment forecast analysis. Although the selection of predictor variables is guided by the research on estimating admission yield at a large public university, the presentation focuses on which method promises greater prediction accuracy and how easily each approach can be explained to enrollment managers who also desire interpretable results.
Description
Keywords
Citation
Extent
1 page
Format
Geographic Location
Time Period
Related To
Table of Contents
Rights
Attribution-NonCommercial-NoDerivs 3.0 United States
Rights Holder
Local Contexts
Email libraryada-l@lists.hawaii.edu if you need this content in ADA-compliant format.