Valuation of Financial Variables in Iranian Companies Using Quasi-Process Correction with LSTM

This study aims to investigate the impact of various financial variables on the prediction of profitability and loss in public companies. Utilizing financial data from an Iranian company listed on the Tehran Stock Exchange, the study applies machine learning techniques, specifically Artificial Neural Networks (ANN), Long Short-Term Memory (LSTM), and Convolutional Neural Networks (CNN), to assess the prediction accuracy of these models. The key financial variables considered include initial asset price, interest rate, volatility, and others. The novelty of this research lies in replacing the LSTM algorithm with ANN, building upon Funahashi's method for analyzing financial data, and providing empirical insights from a real-world dataset. By implementing Quasi-Process Correction (QPC) to minimize errors, the study improves the generalization of the models, enhancing prediction accuracy. The study's results demonstrate the effectiveness of machine learning algorithms in forecasting financial outcomes, with ANN outperforming both LSTM and CNN in terms of minimizing mean squared error (MSE) and mean absolute error (MAE). This research contributes to the field of financial data analysis by offering a comparative evaluation of machine learning models in predicting financial performance, and underscores the importance of accurate predictions for informed investment decisions, risk management, and financial sustainability. The findings also highlight the potential of deep learning techniques to improve the accuracy and reliability of financial forecasts, ultimately supporting better decision-making in capital markets.

Valuation of Financial Variables in Iranian Companies Using Quasi-Process Correction with LSTM Keywords:

Deep learning , Derivatives , Local and stochastic volatility model , CNN , Monte Carlo simulation , LSTM