Predicting Short-Term Stock Price Actions Using Artificial Intelligence

Abstract

Predicting short-term price actions of stocks is a significant aspect of stock market investing. Numerous methods exist, but many lack efficiency and accuracy. With recent advancements in Artificial Intelligence (AI), research has explored its potential in predicting stock markets. However, limited research focuses specifically on short-term price action prediction. This paper discusses the methodology and findings of our AI models in predicting short-term price actions of stocks.We analyzed key metrics from a sample of 80 stocks using various AI models to assess their accuracy in predicting short-term stock prices. The results suggest that AI models can be effective in forecasting short-term stock movements, but their accuracy is highly dependent on the preprocessing techniques and the features used. Notably, the Orthogonal Matching Pursuit CV model achieved the highest accuracy, reaching up to 80%, particularly when previous quarter prices were included in the analysis.