Method and device for predicting traffic data and computer equipment

Abstract

The invention discloses a method and a device for predicting flow data and computer equipment, and relates to the technical field of computer testing, the method comprises: firstly, acquiring historical delivery data and historical traffic data of each historical unit time; based on historical delivery data, calculating at least one input parameter, wherein the input parameters are used for representing characteristics of historical delivery data; based on historical traffic data, calculating at least one reference output parameter, wherein the reference output parameters are used for representing characteristics of the historical traffic data; based on the calculated input parameter and the reference output parameter, training the flow prediction model to obtain a trained flow predictionmodel, and then inputting the acquired planned delivery data in unit time into the pre-trained flow prediction model to obtain predicted traffic data corresponding to the planned delivery data so as to improve the prediction accuracy of the user traffic data.

Description

technical field [0001] The invention relates to the technical field of computer testing, in particular to a method, device and computer equipment for predicting traffic data. Background technique [0002] At present, when the existing vehicles are put on the market, the user flow data in the next preset period is predicted according to the statistical market flow data and advertising cost in the preset delivery period (for example, the delivery period is daily or monthly). For example, when counting independent user data, it is generally based on the increase / decrease ratio of the planned cost of the advertisement to be delivered relative to the cost of the advertisement delivered in the current preset delivery period to determine the relative user traffic in the preset delivery period relative to the The increase and decrease ratio of the corresponding user traffic in the current default delivery cycle, and the relationship between the two is approximately linear. [0003]...

AI Technical Summary

Problems solved by technology

[0003] However, with this traffic forecasting method, there will be a delay in the impact on the traffic after the advertisement is placed. For example, when the delivery is maintained at a high level, the growth rate of the average daily number of unique users will not continue to maintain a high level, and when the delivery drops, The impact on traffic will also be slightly delayed, and due to the existence of marginal effects, the reduced delivery amount during high delivery is weaker than the reduced delivery amount during low delivery. Therefore, based on the corresponding users in the current preset delivery period Traffic, and the increase or decrease ratio of the advertising cost to be delivered in the next preset delivery period relative to the delivered advertisement cost in the current preset delivery period, to predict the user traffic in the next preset delivery period, and the prediction accuracy is low , thus affecting the optimization efficiency of advertising in the market

Images