Search

The intricacies and dynamism of financial markets pose challenges to models seeking to comprehensively capture the multitude of factors influencing stock price movements. As such, there remains room for improvement in forecasting accuracy. In response, we introduce a novel approach that unifies the Root Mean Square Error (RMSE), loss functions of Long Short-Term Memory (LSTM) and Convolutional Neural Networks (CNN). By concurrently optimizing their RMSE loss functions, our novel approach takes use of the capabilities of LSTM for learning long-term time series relationships and CNN for extracting deep features from data. To maximize the efficacy of each model branch within this unified framework, we split the training set into two different representations, one consisting of standard time series data and the other of standard picture data. We compare our proposed model to others in the field to demonstrate its viability, particularly Backpropagation (BP), LSTM, CNN, and a fusion LSTM-CNN model. Experimental evaluations conducted on three diverse datasets—Development Bank, Stock Connect Index (SCI), and Composite Index (CI)—validate the robust predictive performance and applicability of our joint RMSE loss LSTM-CNN model, thus showcasing its potential in financial forecasting. [ABSTRACT FROM AUTHOR]