Federated learning based multi model screening and support system for major depressive disorder

Abstract

Major Depressive Disorder (MDD) is a prevalent mental health disorder characterized by newlinepersistent feelings of sadness, loss of interest or pleasure in daily activities, and a range of newlinecognitive and physical symptoms that impair an individual s functioning. Clinically, MDD newlinemanifests through symptoms such as changes in appetite and sleep patterns, fatigue, difficulty newlineconcentrating, and feelings of worthlessness, often leading to significant distress and disability. newlineThe disorder varies in severity and duration. Given its high prevalence and substantial impact on newlinequality of life, MDD remains a critical focus of mental health research, with ongoing efforts to newlineimprove diagnosis, treatment, and prevention strategies. Automatic screening techniques can be newlineuseful in such situations, as recent advances in Deep Learning (DL) models appear. It showed newlinethat the diagnosis of mental health disorders using DL and machine learning (ML) was effective newlineand suitable for widespread implementation. Since these ML/DL methods share training data newlinewith a centralised server while ignoring ethical and competitive concerns, they are unable to newlineprotect client confidentiality. Federated Learning (FL) is an effort to address the identified newlineshortcomings of centralized training by enabling decentralized training. The depression dataset newlineis gathered from online sources to satisfy the requirement for the dataset to train a DL model. newlineThe commonly used open-source dataset serves as the basis for the research. Before newlineimplementation, every dataset is pre-processed and examined using a number of parameters. newline