Scientific Journal of KubSAU

In present article we consider one of the key problems of short time series in modern scientific prognostification. In scientific papers as a rule the problem of short time series is considered in connection with a problem of small selections. But there are some problems still unsettled what value of selection should be taken as a small one and which time series are short and very short. In spite of the fact that there are exist different opinions on the problem of small selection definition and as a result the definitions of short time series, time series of tax deductions analysis of which is given in the present paper one can qualify as very short ones. The authors are considering tax deductions, the length of which doesn’t exceed twelve meanings. It’s clear that building of the prognostification model on a base of given statistics is impossible. In the present paper the authors offer to carry out analysis of very short time series on the base of graphs and matrixes of similarity. In the base of the method proposed there is a visual method and information base are very short time series of tax deduction and time series of fiscal economic indices, forming taxation base. The given approach should help the officers of tax bodies to fulfill the testing analysis the purpose of which is to select tax-payers for checking up