Detecting psychological symptom patterns using regularized multinomial logistic regression

Salih Tutun, Abdulaziz A. Ahmed, Sedat Irgil, Ilker Yesilkaya, Mohammad T. Khasawneh

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Abstract

    According to the National Institute of Mental Health, for every five adults, one of them lives with a mental disorder such as depression, bipolar disorder, schizophrenia, and anxiety. Such disorders could lead to moderate to severe conditions (i.e., suicide). Psychiatrists use many tools and questionnaires to detect mental disorders and one the most common is Symptom Checklist-90-Revised (SCL-90-R) test, which composed of 90 questions and can detect symptoms for ten different mental disorders. This research proposes a new artificial intelligent framework for detecting mental disorders. The SCL-90-R questionnaire has been sent to more than 500 users in Turkey through a web system and evaluated by a psychiatrist. Then, a prediction model is developed by using Regularized Multinomial Logistic Regression (REM-LR). The results show that the accuracy of the final model is 97%. Experts and users can use the proposed model for enhancing clinical decisions.

    Original languageEnglish (US)
    Title of host publicationIISE Annual Conference and Expo 2019
    PublisherInstitute of Industrial and Systems Engineers, IISE
    ISBN (Electronic)9781713814092
    StatePublished - 2019
    Event2019 Institute of Industrial and Systems Engineers Annual Conference and Expo, IISE 2019 - Orlando, United States
    Duration: May 18 2019May 21 2019

    Publication series

    NameIISE Annual Conference and Expo 2019

    Conference

    Conference2019 Institute of Industrial and Systems Engineers Annual Conference and Expo, IISE 2019
    CountryUnited States
    CityOrlando
    Period5/18/195/21/19

    Keywords

    • Logistic Regression
    • Machine Learning
    • Mental Healthcare
    • Regularization
    • SCL-90-R test

    Fingerprint Dive into the research topics of 'Detecting psychological symptom patterns using regularized multinomial logistic regression'. Together they form a unique fingerprint.

    Cite this