Data Mining Approaches for Depression Detection on Social Media Twitter Dataset

Nowadays, we live in a world where almost everybody has many problems like discomfort and tensions in their lives, irrespective of age, gender, or caste, due to race, customs, and rituals. Due to such problems, some people are struck by inferiority, others take the pressure, and a few become hopeless and try to destroy themselves. This research will utilize several Data Mining techniques, such as association mining, classification, and regression, with Machine Learning algorithms on social media data to extract a user's mental health state, emphasizing depression. The study focuses on diagnosing depression among cross-cultural social media users using association mining of the Data Mining Association and Machine Learning approaches. The proposed methodology has four significant parts discussed in detail in the methodology section. It is concluded that depression is a common but dangerous mental illness. It is a dysfunctional habit that can strike anyone, regardless of age, gender, and socioeconomic background. There are numerous intelligent strategies to spot depression in online life customers. Machine learning algorithms such as Nave Bayes, LSTM, Decision Tree, Logistic Regression, SVM, and Random Forest are used to classify depressed and non-depressed tweets. The mining technique used in this paper is Association Rule Mining; A priori is the best fit to generate the association rules from the tweets dataset and gives the rules for depressed and undepressed tweets with the best performance. More mining techniques and other algorithms could provide more rules and generalization results.