Research on gold price forecasting based on lstm and linear regression

Ocean University of China, Faculty of Information Science and Engineering, Qingdao, 266100, China

^* corresponding author: 20020036017@stu.ouc.edu.cn

Abstract

Gold price forecasting is critical in financial decision-making, providing valuable insights for in-vestors and stakeholders in the gold market. Deep learning methods have witnessed remarkable progress in various domains, including image recognition and sentiment analysis. This paper integrates LSTM (Long Short-Term Memory) and Linear Regression models to forecast the rise and fall of gold prices. The analysis of the prediction accuracy regarding the rise and fall of the daily gold price reveals that the LSTM model achieved an accuracy rate of 50.67%, while the Linear Regression model achieved a slightly higher accuracy rate of 53.02%. By combining the strengths of these models, this research provides valuable insights to investors in the gold markets.