Stock Market Time Series Data Prediction Using Sequence Pattern Mining Algorithms

Abstract

With the continuous improvement of China's financial markets, public investment has gradually shifted toward the stock market. However, stock price fluctuations are influenced by multiple factors, making it difficult for the public to make accurate judgments. To address this, the study proposes a stock market sequence prediction model based on Sequence Pattern Mining algorithms. The model introduces K-line indicator data from the stock market for optimization, forming specific similarity sequences for stocks. By constructing optimal K-line patterns and matching them with stock sequences, the model achieves stock price prediction. Experiments using public datasets show that the proposed model achieves the lowest Mean Squared Error of 18% for K-line data prediction in later iterations. After 500 iterations, the Coefficient of Determination increased from 0.38 to 0.79. The recall rate rose to 88% after 150 iterations. In the analysis of a self-built dataset, the model demonstrated the best predictive performance for the fourth sequence group, with a Mean Absolute Percentage Error of 0.94% and a Root Mean Square Error of 0.95%. Among the prediction accuracy rates for all stocks, the proposed model's accuracy mostly fell within the range of 60%-70%. These results indicate that the proposed model can accurately predict trends in complex financial markets, enhance public confidence in China's stock market, and contribute to the development of the financial industry.