Enhanced stock price prediction with a CNN-BiLSTM deep learning approach optimised by genetic algorithms

People's interest in the stock market has increased in recent years, as economic growth has increased. Accurate predictions of stock price fluctuations provide more financial gain while minimising risk. However, forecasting is challenging due to the constant fluctuations in pricing and their frequently uncertain movements. We propose a hybrid model that combines convolutional neural networks (CNN), bidirectional long short-term memory (BiLSTM) networks, and a genetic algorithm (GA) for stock price prediction. This work compares our model with five other models: CNN, LSTM, BiLSTM, CNN-LSTM, and CNN-BiLSTM. We have used three evaluation metrics to evaluate each model's performance. The proposed GA-CNN-BiLSTM model enhances overall performance by effectively optimising hyperparameters. We analysed a 10-year Nifty 50 dataset and found that the proposed model consistently outperforms baseline models. We conducted additional experiments on global index datasets, including the S&P 500 and Nikkei 225, to assess the model's robustness and significance.