Psychlysis: Towards the Creation of a Questionnaire-based Machine Learning Tool to Analyze States of Mind
Hemakshi Jani, Mitish Karia, Meet Gohil, Rahul Bhadja, Aznam Yacoub, Shafaq Khan
TL;DR
This paper presents Psychlysis, a questionnaire-driven ML approach to analyze current states of mind and deliver personalized mood-enhancement recommendations using the OCEAN personality framework. It situates the work among ML-based mood analysis literature, arguing for a scalable, questionnaire-based method that avoids heavy sensor setups and can assist clinicians and individuals. Preliminary results from 20 participants show moderate applicability of recommendations, while the authors acknowledge data, bias, and multimodal integration challenges requiring broader validation and interdisciplinary collaboration. The work points to future directions including social-media data integration and LLM-enabled chat interfaces to improve questionnaire comprehension and user interaction, all within a strong emphasis on privacy and ethics.
Abstract
This paper describes the development of Psychlysis, a work-in-progress questionnaire-based machine learning application analyzing the user's current state of mind and suggesting ways to improve their mood using Machine Learning. The application utilizes the OCEAN model to understand the user's personality traits and make customized suggestions to enhance their well-being. The proposed application focus on improving the user's mood rather than just detecting their emotions. Preliminary results of the model are presented, showing the potential of the application in predicting the user's mood and providing personalized recommendations. The paper concludes by highlighting the potential benefits of such an application for various societal segments, including doctors, individuals, and mental health organizations, in improving emotional well-being and reducing the negative impact of mental health issues on daily life.