User attribution prediction method, device and equipment based on model adaptive adjustment

Abstract

The invention discloses a user affiliation prediction method, device and equipment based on model adaptive adjustment, and the method comprises the steps: obtaining corresponding prediction basic data according to received event report data, carrying out the binning processing to obtain basic feature data, and carrying out the screening according to prediction configuration information in the event report data to obtain model template parameters, and obtaining sample feature data corresponding to the event report data from the historical database to construct a target prediction model, training the target prediction model according to the sample feature data to obtain model training information, performing attribution prediction on the basic feature data, and obtaining a predicted attribution group corresponding to each user in the event report data. The method belongs to the technical field of intelligent prediction, adaptive configuration of each scene and each industry can be realized by acquiring the model template parameters, and prediction basic data is acquired, so that multi-party data combination is realized, the problem that a prediction result is inaccurate when a general model is used for prediction is solved, and the use requirements of multiple application fields are met.

Description

technical field [0001] The present invention relates to the technical field of intelligent prediction, and in particular to a method, device and equipment for user belonging prediction based on model self-adaptive adjustment. Background technique [0002] In order to obtain the possible changes of the user in the future based on the current behavior of the user, it is possible to predict the user's attribution, that is, to realize the prediction of the user's future attribution group. The existing mainstream user prediction technology basically comes from the lookalike in the advertising field In the field of technology, it is mainly divided into two methods: explicit positioning and implicit positioning. However, the existing user prediction methods are usually only applicable to a specific field. If the application field of the prediction method is expanded, the prediction method will be generalized. Accurate prediction results cannot be obtained in a specific field, so th...

AI Technical Summary

Problems solved by technology

[0003] The embodiment of the present invention provides a user attribution prediction method, device and equipment based on model self-adaptive adjustment, aiming to solve the problem that the user prediction method in the prior art method cannot meet the differentiated use requirements among various application fields The problem

Images