Method and device for predicting information

Abstract

The embodiment of the invention discloses a method and a device for predicting information. A specific embodiment of the method comprises the steps of obtaining features of a user; inputting the features of the user into a pre-trained first model and a pre-trained second model respectively to obtain a first prediction result and a second prediction result of the user, wherein the first model and the second model correspond to different types of mechanisms respectively, and are obtained by performing training by adopting a longitudinal federated learning method based on respective corresponding training samples; and summarizing the first prediction result and the second prediction result to generate a prediction result of the user. The embodiment of the invention relates to the field of cloud computing, and improves the accuracy of information prediction based on the prediction information of the first model and the second model trained by adopting the longitudinal federal learning method.

Description

technical field [0001] The embodiments of the present application relate to the field of computer technologies, and in particular to methods and devices for predicting information. Background technique [0002] Data isolation and island effect seriously restrict the development of artificial intelligence. Different institutions have limited data dimensions and sample sizes. Due to legal restrictions such as data security and personal information protection, data from different institutions cannot be exchanged or shared. As a result, the models established by institutions based on their own data and samples are less effective, have weak generalization capabilities, and are not replicable. [0003] For example, financial institutions have data such as users' deposits, account transaction records, loan amounts, and consumption records, but lack user browsing behavior and interest tags; Internet institutions have users' website browsing behavior, interest tags, and geographic l...

AI Technical Summary

Problems solved by technology

Due to legal restrictions such as data security and personal information protection, data from different institutions cannot be exchanged or shared, resulting in poor model effects, weak generalization ability, and non-reproducibility of institutions based on their own data and samples.

[0003] For example, financial institutions have data such as users' deposits, account transaction records, loan amounts, and consumption records, but lack user browsing behavior and interest tags; Internet institutions have users' website browsing behavior, interest tags, and geographic location information, but lack Financial data owned by financial institutions

Images