Forecasting the Number of Customers on the Basis of Correlation Statistics

Abstract

The paper considers the problem of forecasting the number of customers to service by the example of a car wash. The proposed algorithm is based on finding correlations between the statistical data on the number of clients and amount of daily precipitation. The algorithm consists of three phases: preparation and preliminary processing of the signals, determination of the coefficient of correlation, and predictive analysis. In conclusion, the estimation of the effectiveness of this algorithm is presented. The first stage involves gathering and pre-processing. The stage is based on the application of filters of the lower and upper frequencies, as well as on other numerous methods of signal processing. The second step is based on determining the correlation coefficient and its further use as a measure of forecast deviation. The last step consists of determining the proportions of the signals and applying it to predict the output signal. The use of this processing algorithm and the prediction signal gives employers a much more efficient use of available labor resources using statistical data of clients and amount of precipitation