Mental Disorder Detection via Social Media Mining using Deep Learning

Abstract

Mental disorders are a disease that cannot be physically seen, so diagnosing someone with a mental disorder is not easy. Due to the imperceptible nature of mental disorders, diagnosing a patient with a mental disorder is a challenging task. Therefore, detection in people with mental disorders can be done by looking at the symptoms they experience. One symptom in patients with mental disorders is solitude. Patients with mental disorders feel indifferent to their environment and focus on what is happening to them. Therefore, the patient looks for a place that can accommodate his feelings. Twitter is one of the most widely used media in measuring one's personality through everyday statements. The symptoms as suggested by psychologists can be explored more broadly using Natural Languages Processing. In this case, it takes a lexicon containing keywords that can show symptoms of depression. This study uses five criteria as a measure of mental health in a statement: sentiment, basic emotions, the use of personal pronouns, absolutist words, and negative words. The results show that the use of sentiments, emotions, and negative words in a statement is very influential in determining the level of depression. A depressed person more often uses negative words that indicate his self-despair, prolonged sadness, even suicidal thoughts (e.g. "sadly”, “scared”, “die”, “suicide”). In the classification process, using LSTM Deep Learning obtained an accuracy of 70.89%; precision of 50.24%; recall 70.89%.

Public Opinion Analysis of Presidential Candidate Using Naïve Bayes Method

“…Figure 1 showing the study stage from the data crawling phase to the evaluation phase. The system consists of four main parts covering the processes of data collection, feature extraction, data labeling, and data classification using deep learning [19]. This study used three datasets, namely the Anies Baswedan, Ganjar Pranowo, and Prabowo Subianto datasets.…”

Synthesis of Affective Expressions and Artificial Intelligence to Discover Mental Distress in Online Community

“…In a recent study, authors demonstrated that the use of sentiments, emotions, and negative words in users everyday posts is prominent in determining the level of depression. The classification model was constructed using long short-term memory, a deep learning algorithm which generated an accuracy of 70.89%, precision of 50.24%, and recall 70.89% (Kholifah et al, 2020). In another recent investigation of dataset retrieved from Twitter highlighted posts related to mental risk and by applying multiple instance learning with anaphoric resolution encoder researchers achieved 92% accuracy (Wongkoblap et al, 2021).…”