Mental State Detection From Tweets By Machine Learning

The world over, mental illness is a serious issue. Many people use the social media that may affect their health positively, but often result in negative sentiments. This research aims to determine an individual's state based on behavior Twitter. We analysed dataset including 170000 real tweets by using natural language processing and machine learning techniques. Decision tree, support vector machine, recurrent neural network (RNN) were used for classifying twitter users, detect if they are positive or state. These models compared which approach provides more accurate detection of positive/negative Then, RNN yielded highest accuracy 0.76 among models, with precision, recall, F_1 score being 0.75, 0.74, respectively. truncated singular value decomposition was also utilised visualise high-dimensional feature space data.