Logistic regression and text classification as part of the Yeshiva University Computer Science Capstone Project

Abstract

Text classification is a machine-learning task of assigning labels to a given text. The texts can be categorized based on topics or the sentence’s sentiment like this is a positive or negative sentence. In my capstone project, we were tasked with correctly labeling texts from the Sefaria dataset. The project aims to create a machine-learning model to predict a given Hebrew text’s correct topic or topics. One of my tasks in the project was building logistic regression models for text classification. The results of these models became the initial performance bar to measure the results of other models created in the project. This paper aims to explain what logistic regression is, how to use it, and how it applies to our project.