Depression prediction using machine learning: a review

Hanis Diyana Abdul Rahimapandi, Ruhaila Maskat, Ramli Musa, Norizah Ardi

Abstract


Predicting depression can mitigate tragedies. Numerous works have been proposed so far using machine learning algorithms. This paper reviews publications from online electronic databases from 2016 to 2020 that use machine learning techniques to predict depression. The aim of this study is to identify important variables used in depression prediction, recent depression screening tools adopted, and the latest machine learning algorithms used. This understanding provides researchers with the fundamental components essential to predict depression. Fifteen articles were found relevant. We based our review on the systematic mapping study (SMS) method. Three research questions were answered through this review. We discovered that sixteen variables were deemed important by the literature. Not all of the reviewed literature utilizes depression screening tools in the prediction process. Nevertheless, from the five screening tools discovered, the most frequently used were hospital anxiety and depression scale (HADS) and hamilton depression rating scale (HDRS) for general population, while for literature targeting older population geriatric depression scale (GDS) was often employed. A total of twenty-two machine learning algorithms were identified employed to predict depression and random forest was found to be the most reliable algorithm across the publications.

Keywords


depression; literature review; machine learning; prediction;

Full Text:

PDF


DOI: http://doi.org/10.11591/ijai.v11.i3.pp1108-1118

Refbacks

  • There are currently no refbacks.


View IJAI Stats

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.