THE PREDICTION OF ENTERPRISES ACTIVITY INDICATORS TAKING INTO ACCOUNT THE INITIAL DATA UNCERTAINTY

Abstract

The mathematical method of making forecasts when the initial data uncertainty have been considered. To obtain point and interval estimates of the parameters of mathematical models that take into account errors in the values of the function and argument, the use methods of confluent analysis has been proposed. For linear models, point estimates and confidence intervals of predictions have been found. The developed method is used to solve the problem of forecasting the volume of services provided by organizations of the housing and utilities complex. The sales volume simulating for different values of time intervals has been conducted and the influence of the initial information error on the model parameter values and predicted values has been analyzed.

Keywords

• prediction
• uncertainty
• parameter estimates
• input data
• confidence intervals
• confluent analysis