PSO-tuned support vector classifier based Nifty50 index movement prediction using market positions and sentiment scores

Predicting the movement of the stock price index is a challenging financial time series analysis task. However, an accurate forecast may generate enormous profits for investors. Due to the complexity of stock market data, developing accurate prediction models is arduous. In this context, this study aims to develop an effective model to forecast the direction of daily returns based on the closing price of the Nifty50. The data from multiple social media platforms, including Twitter, StockTwits, Facebook, and YouTube, have been combined for sentiment analysis. The market positions have been used to train the prediction model. Performance comparison of the supervised machine learning models, including support vector classifier and particle swarm optimization-assisted support vector classifier, has been conducted. The model's hyperparameters have been carefully tuned to attain more than 99% prediction accuracy. It has been observed that the novel approach of combining the Nifty50 market positions with social media sentiments resulted in high prediction performance, and the proposed model is robust and dependable to forecast index returns.