This study presents a comprehensive data-driven analysis of the predictive power of financial statement variables on stock returns of ۱۵۰ companies listed on the
Tehran Stock Exchange (TSE) over a ۱۵-year period from ۲۰۱۰ to ۲۰۲۴. Utilizing a dataset consisting of ۹,۰۰۰ quarterly observations, key financial ratios such as
Return on Assets (mean = ۶.۲%, SD = ۳.۸),
Debt-to-Equity ratio (mean = ۰.۸۵, SD = ۰.۵۰), Current Ratio (mean = ۱.۶۵, SD = ۰.۷۰),
Earnings per Share (mean = ۱.۱۲, SD = ۰.۷۵), and Price-to-Book ratio (mean = ۱.۲۰, SD = ۰.۹۰) were examined for their ability to predict quarterly stock returns (mean = ۲.۵%, SD = ۵.۴).The study employs multiple analytical techniques including Multiple Linear Regression (MLR),
Random Forest (RF) regression, and Support Vector Regression (SVR), validated through ۱۰-fold cross-validation to ensure robustness and avoid overfitting. Results indicate that ROA and EPS exhibit statistically significant positive relationships with stock returns (p < ۰.۰۱), whereas
Debt-to-Equity ratio shows a significant negative effect (p < ۰.۰۵).
Random Forest models outperform traditional linear methods by capturing complex nonlinear interactions, improving prediction accuracy by approximately ۱۲%.Additionally, qualitative insights from ۵۰ professional football and sports managers were integrated using thematic analysis, revealing strategic parallels between sports decision-making and financial investment strategies under uncertainty. This interdisciplinary approach offers novel perspectives for investors and managers operating in emerging markets characterized by volatility and regulatory complexity.The findings contribute to the growing literature on financial forecasting and behavioral finance, emphasizing the value of combining quantitative models with domain-specific knowledge. Practical implications include improved investment strategies and risk assessment frameworks for stakeholders in the TSE and comparable emerging markets.
پاسخ به دیدگاه