Stock Price Prediction in Malaysia Retail Industry with First-Order Linear Difference Equation

Abstract

This paper studies the prediction of stock prices in Malaysia's retail industry using a first-order linear difference equation. A retailer's historical stock price data from 1 April 2023 to 1 April 2024 are collected and visualized. A least squares optimization problem is introduced, where the objective function is to minimize the sum of square errors. The first-order necessary conditions are derived, and recursion equations are obtained. The least squares optimization problem is solved using the gradient method, where the model parameters are optimally estimated, giving an optimal solution for the linear model iteratively. Once the convergence is achieved, the optimal solution of the linear model approximates the stock prices closely with a small mean square error value. For illustration, two linear difference equations are considered as the initial models; the first is exponential, and the second is decay. Simulation results show that these initial models provide a satisfactory prediction solution with a small mean square error value. Thus, the predictive model for using different initial models is expressed. In conclusion, the first-order linear difference equation model efficiently predicts the retailer's stock prices. For future research, more complex and high-volatility stock prices can be considered to predict.