Consideration of the dynamics of the daily variance of the index of Mosuri while minimizing risk of short-term operations on the Russian stock market

Abstract

The aim of this work is to develop a model that allows setting a daily interval of the index. The introduction describes the main problems and difficulties associated with forecasting stock markets in General and Russia in particular. In the next part describes in detail the methodology of constructing the model, data preprocessing, parameter selection of autocorrelation. The paper uses such methods as ARIMA, the method of growth indices, the method of artificial neural networks, Fourier spectral analysis, generalization method, two-sample t-test for averages, etc. To assess the results of the study the author used methods of analysis of autocorrelation of the residuals normal distribution of the residuals, the maximum likelihood method, student's t-test, and comparison of forecast based on a row of known (source) data. The evaluation confirmed the hypothesis about the adequacy of the materiality and significance of the model. Based on the model stock strategy is built, designed to minimize investors’ risks. In the discussion so-called "negative results" are given obtained in the course of the study, inadequate models, inappropriate factors, parameters and methods, which had to be abandoned on the way to presented in the result.