Technical Analysis Toolkit for Neural Networks in Finance and Investing

Abstract

Stock market forecasting is an important issue in financial engineering. With the rapid growth of stock market data, artificial intelligence in finance has the potential to provide investors with accurate strategy analysis and effective investment decisions. This study combines the proposed technical analysis toolkit with back propagation neural networks. Many studies discuss how to modify the algorithm of machine learning, but few studies investigate how to achieve good data pre-processing. The proposed technical analysis toolkit provides several useful functions such as stock price analysis, forecasting, and data pre-processing. We use analyzed data rather than raw data as the input variables for the neural networks and verify whether the data pre-processing can provide more accurate stock price predictions. The technical analysis indicator package was written in the R language. The sample data include several major indexes in the U.S. stock market, such as the Dow Jones Industrial Average, Philadelphia Semiconductor Index, Standard & Poor’s 500 Index, and NASDAQ Composite Index. The experiment results indicate that the proposed toolkit helped the neural networks achieve accurate predictions. The proposed toolkit passed a comprehensive R archive network (CRAN) check and contributes to the field of stock data analysis.