Abstract

Mental health is becoming more known among individuals, yet still there is a gap in knowledge and a stigma around it. Depression is a common mental health diagnosis, with around 5% of adults suffering from this. (WHO, 2021). Two weeks of consistent sadness and other factors can lead to depression.

This research uses machine learning to predict signs of depression in young adults in education. Machine Learning is a method which can predict the outcome from data. This research will show how data was collected, inspired by the Depression Anxiety Stress Survey (DASS-21), to find critical variables. From this survey, a questionnaire was created and distributed to the target audience via student platforms and social media to complete.

This research specifically targets young adults between 18-35 in education. Data were pre-processed using Excel and Jupyter Notebook (Python) to work with data to build a suitable prediction model.

A key finding from this research was that life was meaningless and scary without reason with high social media usage. It was noticed that a participant who said they feel like this most of the time had medium or high signs of depression.

A classification model, Random Forest Classifier, proved best out of eight classifiers tested with an accuracy of 79%. Many studies and research have been conducted with uniqueness to their aims and methodology but close in accuracy scores. It will be good to get a larger dataset from the same data collection source, try to improve the accuracy, and launch this application to education establishments.

Using machine learning and a predictive model - Can signs of depression be predicted in young adults within education?