Iranian Financial Engineering Association(IFEA) International Journal of Finance & Managerial Accounting 2588-4379 12 45 2025 03 01 Designing a Model to Predict the Uncertainty of Financial Information by Selecting Features Using Machine Learning and Neural Networks 237 250 24091 10.22034/ijfma.2025.78175.2217 EN Moslem Talebvand Department of Accounting, Borujerd Branch, Islamic Azad University, Borujerd, Iran Mahmood Hematfar Department of Accounting, Borujerd Branch, Islamic Azad University, Borujerd, Iran Nasrollah Takhtaei Department of Accounting, Dezful Branch, Islamic Azad University, Dezful, Iran Journal Article 2024 12 29 This investigation presents a novel approach to designing a model for the uncertainty of financial information in companies listed on the Tehran Stock Exchange. For this purpose, the data of 114 companies listed on the Tehran Stock Exchange are extracted, and the relationship between 12 independent variables and the dependent variable of financial information uncertainty is investigated. Initially, support vector regression (SVR) is used to evaluate the significance of independent variables or select the best features, leading to the selection of five effective independent variables. Then, the artificial neural network (ANN) algorithm with two hidden layers, the random forest algorithm, and the support vector machine algorithm are implemented in Python to analyze and predict the uncertainty level based on the selected features. The analysis results indicate that among the five selected variables, “firm risk,” “ownership structure,” and “earnings smoothing” significantly affect the uncertainty of financial information. Also, based on the mean square error (MSE) criterion, the random forest and neural network algorithms showed very high accuracy in predicting financial uncertainty. This investigation emphasizes the effectiveness of advanced machine learning techniques in predicting financial information uncertainty and provides investors and researchers in the field of finance with invaluable insights. The findings contribute to the development of more accurate prediction models, facilitating risk management and strategic decision-making in a more complex financial prospect. http://www.ijfma.ir/article_24091_9ff6b4c1124dfb0ddeffd72781d88aa6.pdf