THE TECHNIQUE OF ANALYSIS OF MONITORING DATA OF PURCHASE PRICES ON THE EXAMPLE OF THE SARATOV REGION

Abstract

Price monitoring is an important step in the analysis and forecasting of the conjuncture on the grain market. Extremely large or abnormally low price values, which stand out sharply from the general range of other values, have a significant impact on the results of forecasting price movements. The purpose of the article is to substantiate the choice of an effective method for predictive processing of monitoring data that has the property of resistance to the presence of anomalous values in the source data and to a violation of standard assumptions regarding the type of probability distribution, on the basis of which automated data processing can be organized. It is proposed a technique for the on-line analysis of price monitoring data in the grain market, which involves conducting a primary analysis of the laws of probability distribution of time series and applying robust statistical estimates in the case of identifying the deviation from normal (Gaussian) distributions. Using the example of the analysis of procurement prices in the Saratov region, it is shown that the formal application of standard identification procedures and statistical data processing without taking into account their probabilistic characteristics can lead to serious forecasting errors.