Predicting loss aversion behavior with machine-learning methods

SALTIK Ö., ul Rehman W., Soyu R., Degirmen S., ŞENGÖNÜL A.

HUMANITIES & SOCIAL SCIENCES COMMUNICATIONS, vol.10, no.1, 2023 (AHCI, SSCI, Scopus) identifier identifier

  • Publication Type: Article / Article
  • Volume: 10 Issue: 1
  • Publication Date: 2023
  • Doi Number: 10.1057/s41599-023-01620-2
  • Journal Name: HUMANITIES & SOCIAL SCIENCES COMMUNICATIONS
  • Journal Indexes: Arts and Humanities Citation Index (AHCI), Social Sciences Citation Index (SSCI), Scopus, Index Islamicus, Directory of Open Access Journals
  • Dokuz Eylül University Affiliated: No

Abstract

This paper proposes to forecast an important cognitive phenomenon called the Loss Aversion Bias via Hybrid Machine Learning Models. One of the unique aspects of this study is using the reaction time (milliseconds), psychological factors (self-confidence scale, Beck's hopelessness scale, loss-aversion), and personality traits (financial literacy scales, socio-demographic features) as features in classification and regression methods. We found that Random Forest was superior to other algorithms, and when the positive spread ratio (between gain and loss) converged to default loss aversion level, decision-makers minimize their decision duration while gambling, we named this phenomenon as "irresistible impulse of gambling".