نوع مقاله : علمی ـ پژوهشی
نویسندگان
1 دانشجوی دکتری اقتصاد مالی دانشگاه شهید اشرفی اصفهانی، اصفهان، ایران
2 هیات علمی دانشگاه شهید اشرفی اصفهانی ، اصفهان، ایران
3 عضو هیات علمی و استادیار دانشکده معارف و اندیشه اسلامی دانشگاه تهران
کلیدواژهها
عنوان مقاله English
نویسندگان English
. Cognitive biases can lead to misinterpretation of information, flawed analyses, and ultimately detrimental financial choices. The present study aims to identify and analyze these biases in the Iranian stock market, employing machine learning algorithms to model investor behavior. Data related to 10 selected stocks from the Iranian market were collected over a specified time period, preprocessed, and prepared for training machine learning models. In this research, two algorithms—logistic regression and random forest—were applied for performance comparison. The results indicated that the random forest model, with an accuracy of 97%, significantly outperformed the logistic regression model, which achieved 71% accuracy. These findings demonstrate the higher effectiveness of more sophisticated algorithms in understanding and analyzing investors’ irrational behaviors. The study was conducted without any bias in model selection, and all results are presented transparently and reproducibly. The primary objective is to enhance rational decision-making in the capital market through data-driven analysis and artificial intelligence tools. Finally, we demonstrate that each identified cognitive bias manifests in specific behavioral patterns in the Iranian financial market, influenced by the cultural roots of the population.
کلیدواژهها English
