Comparative analysis of methods based on moving average and singular decomposition in stock prices behavior prediction

Abstract

In the article the task of forecasting the price behavior of a stock based on the observation of the realizations of the price of the securities and the criteria for their optimal periods for buy or save are formally determined for the first time. Using the data of the ETF's stock exchanges and the information technology developed for their analysis, a comparative analysis of methods based on the moving average and singular expansion of stochastic stock quotes data was conducted. The accuracy of these forecasting methods is compared. It has been determined that moving-average indicators have a greater likelihood of making the right buy or sale decision. The scientific and applied direction of further research is defined: the analysis of percentage changes instead of absolute values of prices, the division of the studied series into two categories: with the general tendency to growth and decline and conducting two independent experiments.