Cuckoo Search Optimization and Long-Term Memory-based Stock Price Prediction Model with Improved Classification Performance

Academics and business people alike are finding it increasingly difficult to accurately anticipate stock prices. An important indicator of future stock prices is the mood of investors. Users may now voice their opinions on a wide range of firms’ stock and discuss their predictions for the stock market’s future thanks to the proliferation of social media and online forums. In order to accurately predict stock prices, it is helpful to do a sentiment analysis (SA) on all available data pertaining to equities, as well as their historical prices. For this reason, this research introduces a unique cuckoo search optimization (CSO) for stock prediction utilizing Twitter data, with a SA based on Long Short-Term Memory (LSTM). The suggested CSO-LSTM method utilizes an analysis of Twitter data to provide accurate stock price forecasts. When the CSO method is used to fine-tune the LSTM model’s parameters, the model’s classification performance is also considerably improved. Extensive simulations are run with Twitter data, and experimental findings demonstrate that the CSO-LSTM methodology outperforms state-of-the-art methods.