Effect evaluation method and device for information push, electronic equipment and storage medium

Abstract

The invention provides an information pushing effect evaluation method and device, electronic equipment and a storage medium. The method comprises the steps of determining a feature vector of to-be-pushed information; inputting the feature vector into an effect evaluation model to obtain an effect evaluation result output by the effect evaluation model, the effect evaluation model being obtained based on a sample feature vector of to-be-pushed sample information, a sample fact output vector and effect label training; the effect evaluation model is constructed on the basis of the generative adversarial network, the composition of the model is simplified on the basis of a traditional scheme, so that the lightweight effect evaluation model is obtained, meanwhile, input parameters of the model can be greatly reduced, various problems caused by the fact that two adversarial generators are needed in the traditional scheme are completely solved, and the method is suitable for large-scale popularization and application. According to the method, lightweight calculation is realized, namely, the effect evaluation process is faster, real-time decision can be supported, and the method has relatively high implementability.

Description

technical field [0001] The present invention relates to the technical field of data processing, and in particular, to a method, device, electronic device and storage medium for evaluating the effect of information push. Background technique [0002] Although A / B test can better evaluate the effect of intervention, it involves experimental grouping and requires intervention or no intervention on different experimental groups (control experimental group and control experimental group) to obtain control experimental group and control experimental group. Different groups of experimental effects. [0003] However, in practical scenarios, the implementation of randomized controlled trials not only requires a lot of time and effort, but also has high economic costs, and the controlled trials are also limited by various factors, which make their feasibility not high, such as news feeds. When the company is designing products, it needs to evaluate the impact of device types (mobile ...

AI Technical Summary

Problems solved by technology

[0004] The present invention provides an effect evaluation method, device, electronic equipment and storage medium for information push, which are used to solve the defects in the prior art that the cost of the effect evaluation scheme is too large, the operation process is complicated, and the parameters are redundant

Images