A method and system for traffic forecasting

Abstract

The present invention provides a method and system for forecasting traffic volume, which helps to avoid the influence of human factors on traffic volume forecasting, thereby improving the accuracy and stability of traffic volume forecasting. The method includes: collecting the historical traffic volume according to the collection time interval in the collection period; forming the historical traffic volume into a matrix with the collection date as the row and the collection time point as the column, performing singular value decomposition on the matrix, and Obtain an orthogonal matrix with the collection date as the row and the collection date as the column; use the linear regression model obtained by modeling the historical traffic volume to obtain the forecasted traffic volume on the forecast day according to the traffic volume of the current day, The forecast day is the next day of the current day, and the linear regression model is to add the product of the first regression parameter and the traffic volume of the current day to the second regression parameter to obtain the forecast traffic volume of the forecast day; output the The predicted traffic volume on the forecast day.

Description

technical field [0001] The present invention relates to the technical field of computers and software, and in particular, to a method and system for predicting traffic volume. Background technique [0002] In the process of enterprise-to-customer service, the telephone service of the call center is an important service method. Its main responsibility is to be responsible for customer consultation, after-sales and other business problems when customers call, and arrange customer service personnel to answer the phone and solve customer problems. [0003] Since the work arrangement of the customer service personnel mainly depends on the daily traffic volume, it is necessary to make a forecast on the traffic volume. Traffic volume prediction can help the customer service department of the enterprise to schedule shifts, avoiding the waste of manpower with less call volume but more idle customer service staff, or avoiding the occurrence of excessive traffic volume but insufficien...

AI Technical Summary

Problems solved by technology

[0008] The disadvantages of prior art methods are: relying on subjective factors of expert experience and low prediction accuracy and poor stability

Affected by experts' understanding of business and personal experience, the accuracy of predictions cannot meet service expectations

Images