Research on Market Trend Prediction Algorithm Based on High-Frequency Trading Big Data Analysis and Pattern Recognition

This paper studies the problem of market trend prediction based on high-frequency trading big data and proposes an algorithm that combines pattern recognition with the autoregressive integrated moving average model (ARIMA). This paper designs a market trend prediction algorithm based on historical trading data, uses pattern recognition to analyze the potential price fluctuation rules in the market, and combines the ARIMA model to accurately predict market trends. The core steps of the algorithm include data preprocessing, feature extraction, model training and result prediction. Through simulation tests on large-scale high-frequency trading data, this paper verifies the effectiveness of the designed algorithm in an actual trading environment. The simulation results show that the prediction method based on pattern recognition can significantly improve the prediction accuracy of market trends. Compared with traditional methods, the prediction error is reduced by about 10%, which has significant advantages in capturing short-term market fluctuation trends. The research results of this paper provide new ideas for market trend analysis and prediction in high-frequency trading, and have wide practical application value.