Analysis of Therapy Transcripts usingNatural Language Processing
Sarah Hawa1, Shriya Akella2, Shrishti Kaushik3, Vrushali Joshi4, Dhananjay Kalbande5
1Sarah Hawa*, Department of Electronics and Telecommunication, Sardar Patel Institute of Technology, Mumbai, India.
2Shriya Akella, Department of Electronics and Telecommunication, Sardar Patel Institute of Technology, Mumbai, India.
3Shrishti Kaushik, Department of Electronics and Telecommunication, Sardar Patel Institute of Technology, Mumbai, India.
4Vrushali Joshi, Department of Electronics and Telecommunication, Sardar Patel Institute of Technology, Mumbai, India.
5Dhananjay Kalbande, Department of Computer Science, Sardar Patel Institute of Technology, Mumbai, India.
Manuscript received on August 07, 2020. | Revised Manuscript received on August 15, 2020. | Manuscript published on August 30, 2020. | PP: 489-494 | Volume-9 Issue-6, August 2020. | Retrieval Number: F1598089620/2020©BEIESP | DOI: 10.35940/ijeat.F1598.089620
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Mental health plays an integral part in leading a healthy life and having a positive outlook. This impacts our behavior, thought process, and actions and therefore it’simportant to identify and detect mental disorders in an early stage as it’s effects can have a lasting influence on one’s life. According to WHO, one in four people get affected by mental health disorders and currently 450 million people suffer from such conditions. Natural Language Processing can be a useful tool to analyze the trends in therapy transcripts. They can be further trained and optimized to derive useful insights and predict plausible future trends. Our proposed system analyses therapy transcripts and classifies it as ’Early signs of depression’ and ’Serious after-effects of prolonged depression’ based on the nature of the responses. Our system uses three different classifiers- Naïve Bayes, Support Vector Machine, and Logistic regression as well as two different victories- TF-IDF and Count, to classify the text into these categories. This proposed system will not only help patients in identifying their symptoms but will also help therapists and researchers in gathering a large amount of data which could be used in predictive analysis, diagnosis and understanding the patient. Such research will pave the way for improving counselling and therapy sessions and be a very essential analysis tool for therapists.
Keywords: Naïve Bayes, TF-IDF, Vector Machine.