Stock Price Prediction Based on Trend Characterization

In the current economic landscape, stocks have evolved into an essential vehicle for investment and wealth management, with stock trading constituting a vital facet of modern economic activity. Fluctuations in stock prices directly impact investor returns, and the ability to forecast stock price changes could minimize investment risks and augment returns. Consequently, stock price prediction has emerged as a focal point of research. The evolution of machine learning technology has led to the progressive application of these techniques in stock investment decision-making. This study introduces an enhanced KNN model for predicting stock prices. Grounded in the basic principle of KNN, this model restructures the input feature attributes by linking continuous multi-day trading indicators to create a short-term price change trend. This trend is then used as the KNN model's input to predict subsequent trading days' stock prices. Experimental trials indicate that the enhanced KNN model outperforms the comparative algorithms in predictive performance.