Research Article
Student Depression Prediction Using Feature Selection and Logistic Regression Algorithm
@INPROCEEDINGS{10.4108/eai.19-12-2025.2362524, author={Dileep Kumar Sharma and Aaryaa Khatiwada and Tanisha Thapa and Manaswee Niraula and Abhilasha Sharma}, title={Student Depression Prediction Using Feature Selection and Logistic Regression Algorithm}, proceedings={Proceedings of the 1st International Conference on Computational Techniques in Data Science, IICTDS 2025, 19-20 December 2025, Chandigarh, India}, publisher={EAI}, proceedings_a={IICTDS}, year={2026}, month={11}, keywords={Depression prediction logistic regression data pre-processing machine learning feature selection}, doi={10.4108/eai.19-12-2025.2362524} }- Dileep Kumar Sharma
Aaryaa Khatiwada
Tanisha Thapa
Manaswee Niraula
Abhilasha Sharma
Year: 2026
Student Depression Prediction Using Feature Selection and Logistic Regression Algorithm
IICTDS
EAI
DOI: 10.4108/eai.19-12-2025.2362524
Abstract
This study applies logistic regression to predict depression among students in various classes, including 12th grade, undergraduate, postgraduate, and PhD courses, within a specific age group. Initially, a dataset consisting of 27901 records, having 17 input features and a target feature, is acquired. Different techniques such as missing value handling, level encoding, standardisation and dataset balancing using random oversampling are employed to preprocess the dataset. In the next step to minimise the problem of overmodelling and computational complexity, 13 important features are selected using random forest as a feature selector. Hyperparameter tuning using grid search has been performed to select the regularisation parameter of the regression model. The selected features are used to train the logistic regression classifier. The efficacy of the algorithm is tested with four different models. It is seen that the suggested method excels other classifiers in relation to recommended parameters to classify the scholars under depression.