Stock Price Prediction with Fine Tuning of Support Vector Machine Using Sea Lion Optimization

One of the most prominent exercises people do is foresee the stock market. While attempting to predict an exact price of the stock is a futile attempt, we can accurately identify the direction, in which the stock price will move. A large number of the elements that influence the stock market come in the form of news items. Because traders are quickly swayed by what they read, the sentiments that users derive from news headlines have a significant impact on their purchasing and selling behaviors. Our research presents a hybrid technique that integrates the Support Vector Machine (SVM) and Sea Lion optimization algorithm (SLnO). Hyperparameters (kernel function, Regularization parameter, Kernel coefficient) of SVM are optimized using SLnO. Optimized SVM is employed to foresee the intraday close price movement. We have used Nifty 50 historical data from Yahoo Finance and sentiment scores from news headlines aggregated from Pulse (Pulse by Zerodha,2020) to assess the proposed hybrid strategy. According to the experimental findings, the hybrid SVM and SLnO model performs better than the benchmark model.