Algorithms For Predicting Student Dropouts With Feature Selection

Abstract

Proof that a feature is applicable has become a prerequisite for using data mining calculations successfully in real-world contexts. In order to achieve the relevant feature subsets in the writing's order and grouping goals, numerous feature selection approaches have been offered. The concepts of feature pertinence, general strategies, assessment standards, and the qualities of feature selection are presented in this work. Last but not least, the chi square test will be used in feature selection calculations to predict school dropouts. This paper's goal is to find comparable instances of usage in the data compiled from datasets so that, in the end, we may make predictions for each student based on various segment, scholastic, and point-of-view features. In conclusion, information gleaned from the study may shed light on how to support students who are in risk even more effectively. We will wrap up this work with real-world applications (such early student dropout prediction), challenges, and potential directions for future study.