Predicting Depression and Sleep Issues from Digital Behavior Using Machine Learning

The use of social media, screentime, and phone usage has been associated with mental health challenges and negative impacts on sleep. Predicting these declines and impacts can help individuals become more aware of harmful digital habits and possibly adjust them. In this paper, models were built to predict depression and sleep issue frequency based on a number of demographic and behavioral variables. These variables were obtained from a dataset consisting of survey responses from 481 respondents asking about mental health and behavioral indicators including frequency of feeling worried or depressed. The models included Logistic Regression, Decision Tree, Random Forest, Gradient Boosting, K-Nearest Neighbors, Neural Networks, and Convolutional Neural Networks. The best performing models overall in terms of predicting depression frequency were K-Nearest Neighbors, Random Forest, and Logistic Regression. As for predicting sleep issues, Logistic Regression achieved the highest accuracy before tuning at 71%, while Neural Networks performed best after tuning at 68%. The results of this research may help identify patterns between digital behaviors and self-reported mental health outcomes, providing a foundation for further research.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-15
DOI
https://doi.org/10.5281/zenodo.24265033
Primary Topic
Artificial Intelligence in Healthcare
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Predicting Depression and Sleep Issues from Digital Behavior Using Machine Learning

Paul Lieber
Zenodo (CERN European Organization for Nuclear Research)
Artificial Intelligence in Healthcare
preprint

Predicting Depression and Sleep Issues from Digital Behavior Using Machine Learning

Paul Lieber
preprint en

Abstract

The use of social media, screentime, and phone usage has been associated with mental health challenges and negative impacts on sleep. Predicting these declines and impacts can help individuals become more aware of harmful digital habits and possibly adjust them. In this paper, models were built to predict depression and sleep issue frequency based on a number of demographic and behavioral variables. These variables were obtained from a dataset consisting of survey responses from 481 respondents asking about mental health and behavioral indicators including frequency of feeling worried or depressed. The models included Logistic Regression, Decision Tree, Random Forest, Gradient Boosting, K-Nearest Neighbors, Neural Networks, and Convolutional Neural Networks. The best performing models overall in terms of predicting depression frequency were K-Nearest Neighbors, Random Forest, and Logistic Regression. As for predicting sleep issues, Logistic Regression achieved the highest accuracy before tuning at 71%, while Neural Networks performed best after tuning at 68%. The results of this research may help identify patterns between digital behaviors and self-reported mental health outcomes, providing a foundation for further research.

Zenodo (CERN European Organization for Nuclear Research)
Artificial Intelligence in Healthcare
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon