Investigation of financial time series by means of multidimensional statistical analysis

Abstract

The article analyzes the relationship of financial time series using methods of multidimensional statistical analysis. The presence and nature of the relationship between stock indices of different countries is shown. These are the Russian MICEX index, the Chinese SZSE, the American S&P 500, the FTSE 100 stock index of the UK, the German DAX, the BSE Sensex 30 stock index of India, the DFM General index of the UAE and the French CAC 40. Regressions are constructed, where the MICEX is taken as a dependent variable, both for all indices together and separately. The indices were also grouped using factor analysis, the analyses of which were compared during rotation and cluster division. Also, through cluster analysis, periods defined by special characteristics were identified. Regression analysis was also carried out in each cluster and certain patterns of behavior and the degree of influence of the selected indices were derived.