Ensemble Hybrid Learning Methods for Automated Depression Detection

Abstract

Changes in human lifestyle have led to an increase in the number of people suffering from depression over the past century. Although in recent years, rates of diagnosing mental illness have improved, many cases remain undetected. Automated detection methods can help identify depressed or individuals at risk. An understanding of depression detection requires effective feature representation and analysis of language use. In this article, text classifiers are trained for depression detection. The key objective is to improve depression detection performance by examining and comparing two sets of methods: hybrid and ensemble. The results show that ensemble models outperform the hybrid model classification results. The strength and effectiveness of the combined features demonstrate that better performance can be achieved by multiple feature combinations and proper feature selection.

Detection and prediction of Future Mental disorder from Social Media Data using Machine Learning, Ensemble Learning, and Large Language Models.

“…The approaches proposed in this work and the state-of-the-art approach used the same REDDIT to train and evaluate the suggested models. Kumar et al [8] 2019 100 users 1 (depression) 3 ML models Acc= 85% Tarik et al [9] 2019 Not mentioned 1 (depression) 2 DL models Acc= 74% Hussain et al [10] 2019 Not mentioned 2 (depression & anxiety) 3 ML models F1-score= 0.84 Wong et al [11] 2019 Not mentioned 1 (depression) 2 DL models -Rezaii et al [12] 2019 40 users 1 (depression) 2 NLP techniques Acc= 90% Inkpen et al [13] 2019 Not mentioned 2 (depression and PTSD) 1 DL model Acc= 88% Thorstad et al [6] 2019 All REDDIT 4 (all mental disorders) 1 ML model F1-score= 0.77 Trifan et al [14] 2020 Not mentioned 1 (depression) 3 ML models F1-score= 0.72 Jiang et al [15] 2020 Not mentioned 4 (all mental disorders) 1 DL model F1-score= 0.64 Alghamdi et al [16] 2020 Not mentioned 1 (depression) 6 ML models Acc= 80% Birnbaum et al [17] 2020 223 users 1 (depression) 2 ML models Acc= 77% Chatterjee et al [18] 2021 Not mentioned 1 (depression) 1 ML models Acc= 76% Ren et al [19] 2021 Not mentioned 1 (depression) 1 DL models Acc= 91% Shaoxiong et al [20] 2022 All REDDIT 1 (depression) 2 EL models Acc= 75% Nalini. L [21] 2022 Not mentioned Not mentioned 3 ML models Not mentioned Tufail [22] 2023 Not mentioned 1 (depression) 1 DL model Acc= 64% Koushik et al [23] 2023 Not mentioned 1 (depression) 1 ML & 2 DL Acc= 60% Yicheng et al [25] 2023 Not mentioned 1 (depression) 1 Time series approach -Helmy et al [27] 2024 70,000 tweets 1 (depression) 5 ML models Acc=92% Dhariwal [28] 2024 Small healthcare dataset The proposed work aims at the early detection and even the prediction of potential future mental disorder from social media data.…”

“…Afterwards, in 2022 Shaoxiong et al [20] trained many classifiers on text to detect depression. The study mainly focused on the comparison between hybrid and ensemble methods in automatic depression detection from social media posts.…”

Deep learning and machine learning in psychiatry: a survey of current progress in depression detection, diagnosis and treatment

“…From this table we see a heavy reliance on text data. Recently, we observe a trend away from hand-crafted features towards complex neural word embedding models such as those seen in [ 59 , 58 , 62 ]. This mirrors a pattern seen in the data science field in general with powerful text embedding models becoming the current state of the art.…”

Hybrid Approach for Enhanced Depression Detection using Learning Techniques

“…Recently, 'Sushant Singh Rajput' attempted suicide due to depression [1]. The goal of this system is to create a hybrid modal system [9] that will help society and medical professionals diagnose and track depression. This study utilized ensemble learning to learn from PHQ8 data in order to get the correct positive results [30].…”