Stock price behavior determination using an optimized radial basis function

Abstract

Multiple stock exchanges make up the global stock market. Shares are purchased and sold by the public and investors, and their prices fluctuate in accordance with supply and demand. Stock trading is one of the simplest investment strategies due to its widespread use in finance. The intuitive ability of investors to make more informed investment decisions should be facilitated by the accurate prediction of a stock price. However, the stock market is highly complex, nonlinear, and non-stationary due to the multitude of factors that influence individual stock prices. Given this circumstance, it is exceedingly challenging to accurately forecast stock prices. Necessity for machine learning was derived from the challenges and low accuracy of conventional techniques. The radial basis function (RBF) is a type of machine learning model that is particularly effective at predicting changes in the stock market. This model is combined with the battle royale optimizer (BRO) to generate the most precise prediction results for the Shanghai Stock Exchange index (SSE). Complex data sets can be approximated by these functions. This is particularly appropriate for the processing of the large volumes of data that the stock market forecast necessitates. The RBF model was combined with the BRO to achieve the most optimal performance in forecasting the stock market for the SSE index. This study examined market trends from January 5, 2015, to June 29, 2023. Additionally, the robustness and ability of the proposed mode were evaluated by using 5-fold cross-validation and statistical tests. A variety of benchmark models that demonstrate the effectiveness of the suggested model on other indexes, such as the FTSE 100, Dow Jones, and HSI, have been used to demonstrate its efficiency. The proposed model was, in fact, highly reliable in its ability to predict market trends, rendering it extremely beneficial to both investors and traders.