Evaluation of Deep Learning Models for Time Series Forecasting in Financial Markets

Abstract Stock value prediction is a challenging task due to unstable and non-lectured nature of financial markets. This research examines the application of long-term short-term memory (LSTM) and Convolutional Neural Network (CNN) in forecasting stock price by taking advantage of historical market data. This study compares the effectiveness of these models in predicting stock price movements, using key matrix such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), R-squared (R²) and accuracy. Integrating LSTM and CNN increase both temporary dependence and the future accuracy by capturing important market trends.