Talk to Me: Artificial Intelligence “Virtual Friend” for Depression Sufferers Using Term Frequency-Inverse Document Frequency (TF-IDF) and Finite State Machine Method

Abstract

Depression refers to mental health in which a person experiences a bad mood and has a decreased quality of life. In Indonesia, there are quite a lot of challenges in dealing with depression problems such as lack of education on how to treat depression, lack of mental health personnel, and the emergence of a bad stigma against depression sufferers. Voice-based artificial intelligence technology for people with depression was developed to play a role in filling the gap by acting as a support system. In this research, the Natural Language Processing (NLP) method is used to enable computer to be able to understand the user's input. TF-IDF (The Term Frequency-Inverse Document Frequency) method is also used to weight documents and the Finite State Machine (FSM) method used to classify the results of document weighting against a predetermined dialogue scenario. To be able to interact with the system, the author uses the Google Cloud Speech API technology to convert speech and text. As for testing of this system, it is done by calculating the level of accuracy of the answers given by the system to users. The level of accuracy of the system answers obtained from the test results is 96.5%. The accuracy value indicates that the answer given by the system is in accordance with what the user's input.