Clarification of stock market structure using wavelet analysis

Abstract

Although predicting stock price by machine learning is an important theme both in theacademic and business fields, it is also extremely important to elucidate the structure of the backgroundstock market behind the predict model. In this research, we aim to elucidate what kind of factors are in thebackground of the stock price prediction by using wavelet analysis. From the historical stock pricetechnical index and various financial market data, we create a model to predict the sign of the waveletcoefficient. We also visualize the mechanism of stock price forecasting by elucidating in time series thetypes of indicators contributing greatly in the prediction model.