User loss prediction method and system

Abstract

The invention discloses a user loss prediction method and system, and relates to the technical field of networks. The method comprises the following steps of acquiring basic information indexes and behavior information indexes of users from a server; selecting part of the users from the users acquired in a statistics period to serve as sampled users, and selecting target indexes and M indexes of the sampled users; building a potential lost user identification model through a decision tree algorithm according to the target indexes and the M indexes of the sampled users; and taking values of the M indexes acquired after the statistics period as input variables of the potential lost user identification model, obtaining a loss probability of the users corresponding to the M indexes, and if the loss probability is greater than a set threshold, judging that the users are potential lost users. According to the method and the system, through the potential lost user identification model built based on the selected basic information indexes and behavior information indexes of the users, the model solidification is simple and convenient, the modeling efficiency is high, and the potential lost users can be accurately identified.

Description

[0001] The present invention relates to the field of network technology, in particular to a user loss prediction method and system. Background technique [0002] With the rapid development of the live broadcast industry, the competition among various live broadcast platforms has become increasingly fierce, leading to severe user loss and causing direct economic losses to the live broadcast platform, which is not conducive to the long-term development of the platform. Therefore, it is necessary to analyze user behavior through machine learning algorithms, establish a potential lost user identification model, accurately locate potential lost users, and formulate maintenance strategies to restore potential lost users. At present, user behavior tends to be diversified and complicated, and changes rapidly, resulting in a large number of user behavior information indicators. If the user's behavior parameters are directly used as the input variables of the model, the modeling process ...

AI Technical Summary

Problems solved by technology

At present, user behavior tends to be diversified and complicated, and changes rapidly, resulting in a large number of user behavior information indicators. If the user's behavior parameters are directly used as the input variables of the model, the modeling process and the model itself will be more complicated, and The model is not very accurate in predicting churn

Images