Stock Price Forecasting Based on the Gauss-Newton Method

Abstract

The purpose of this study is to explore the stock price forecasting method based on the Gauss-Newton method. In this work, through a literature review, the existing stock forecasting methods were analyzed, and it was noted that there is a lack of attempts to construct nonlinear regression models using the Gauss-Newton method in existing research. The practical operation section provides a detailed description of data reading, fitting of the logarithmic regression model, the construction process of the exponential regression model and the power function regression model, as well as parameter estimation and forecasting using the Gauss-Newton iterative method. By comparing the mean square error and graphical analysis of different models, the optimal model is determined. The significance of this study lies in providing investors with a new tool for understanding market fluctuations, which helps in formulating more scientific risk management strategies.