چکیده مقاله
The evolution of relative valuation in finance reflects a continual search for more accurate and theoretically grounded methods to estimate firm value Early valuation practices, pioneered by Graham and Dodd in the 1950s, relied on simple financial multiples such as price to earnings and price to book ratios Although these models provided intuitive comparability, their linear and static nature limited predictive accuracy in increasingly complex markets Subsequent advancements by Damodaran and others formalized the use of multiples and refined comparability standards, yet traditional regression based approaches remained constrained by their inability to capture nonlinear relationships and multidimensional dependencies The emergence of machine learning ML has redefined the relative valuation landscape, introducing models capable of learning intricate patterns among firm fundamentals, market indicators, and macroeconomic variables This paper traces the historical and methodological development of relative valuation methods, comparing classical and ML based frameworks in terms of accuracy, interpretability, and data requirements The findings highlight that ML enhances valuation precision by integrating nonlinear modeling and data driven feature selection while maintaining theoretical coherence through explainable AI XAI tools such as SHAP and LIME The study concludes that the future of valuation lies in hybrid systems that combine the conceptual rigor of traditional finance with the adaptability and transparency of modern ML models, thereby establishing a more robust and explainable foundation for firm valuation in data driven financial markets
کلیدواژهها
نویسندگان
Bahram Rezvani
Msc student of Information System Management - Advanced Information Systems, College of Management, University of Tehran, Tehran, Iran
Zahra Sadat Mirhadi Shirazi
Department of Finance, Khatam University, Tehran, Iran
شیوه ارجاع
Mirhadi Shirazi, Zahra Sadat,1404,The Evolution of Relative Valuation: From Financial Multiples to Machine Learning Models,The 11th international conference on key researches in management, accounting, banking and economics,Mashhad
ارائهشده در
مجموعه مقالات یازدهمین همایش بین المللی پژوهش های شاخص در مدیریت، حسابداری، بانکداری و اقتصاد28 آبان 1404 · مشهد