Approach Based on Ontology and Machine Learning for Identifying Causes Affecting Personality Disorder Disease on Twitter

This paper presents an approach for the detection of the reason that causes a personality disorder (PD) among Twitter users. A reason is the subject of a tweet that is recorded with an inappropriate mood by an individual having already a PD. For this reason, we will focus on our work on the identification of the tweet's topic and the mood of the tweet's writer. Our approach includes a step to cluster topics based on an unsupervised recursive learning method. Besides, a step for elaborating for each leaf obtained from the previous step an ontology which will be used later for decision making by querying it using the different rules obtained automatically. Our proposed method benefits from both natural language processing (NLP) techniques and artificial intelligence (AI) algorithms. These techniques gave us the opportunity to obtain an explanatory result in the form of a hierarchical tree destined for an open lexicon. As an evaluation result, we got an F-score equals to 73% for the evaluation of topic identification and 92% for mood detection.