STATISTICAL ANALYSIS OF VARIABILITY OF DATA OF THE AGRARIAN SECTOR OF KAZAKHSTAN ECONOMY

The article studies the variability of a data group formed from indicators of the agricultural sector of the economy of Kazakhstan, using one of the methods of reducing dimensionality as a factor analysis. The data of 1993–2018, i.e. for 26 years. The analysis included: a) indicators of GDP - the agrarian sector of the economy as a whole and by crop and livestock industries; b) indicators of productivity and gross harvest by type of crop. Statistical processing of the source data was performed using the STATISTICA_10 software package. The results of statistical processing of primary data - GDP indicators: a) fluctuations in GDP data, estimated by the coefficient of variation, were in the range of 65-67%; b) according to the values of pair correlation coefficients (0,97 ... 0,99), it can be argued that the considered GDP indicators are interdependent variables with a close correlation. It was found that almost 90% of the variance of the variables, estimated by their total dispersion, is a consequence of the influence of the three identified factors. The first factor explaining the variability of the main part of the variables, in our opinion, can be taken as the «specific gravity» of the influence of natural and climatic conditions. Close correlation relationships have been established (correlation coefficients from -0,72 to -0,96) between the first factor and all variables of the «productivity» group. The influence of the first factor is also noticeable in relation to GDP indicators, this can be judged by the values of the correlation coefficients: -0,85; -0,85; -0,85. Other identified factors show the level of influence of other objective reasons that are not yet amenable to logical explanation