Insurance product recommendation method and system

Abstract

The invention provides an insurance product recommendation method and system, and the method comprises the steps: S1, obtaining a plurality of insurance data from a multi-dimensional data source, including the personal dimension information of a user and the feature attributes of an insurance product; s2, preprocessing and integrating the insurance data to obtain a plurality of integrated information; s3, presetting and using a plurality of machine learning algorithms for training, and selecting one algorithm from the integrated information of different data sources as a training model in combination with a loss function to obtain a plurality of recommendation models; s4, performing weight calculation on each recommendation model, and performing ensemble learning by using multiple recommendation models to obtain an ensemble recommendation model; s5: according to the user personal dimension information input by any user, predicting a plurality of recommended feature attributes by utilizing an integrated recommendation model; and S6, matching a plurality of insurance products from an insurance product library by utilizing the characteristic attributes, and recommending the insuranceproducts. According to the invention, the insurance data is obtained from the multi-dimensional data source, and the recommendation strength of insurance products is improved through the insurance data.

Description

technical field [0001] The invention relates to the field of information technology, in particular to an insurance product recommendation method and system. Background technique [0002] The insurance awareness of existing users is getting stronger and stronger, and more users are beginning to actively purchase insurance. There are many types of insurance in the market, and it takes a long time for users to choose an insurance product that suits them from these insurance products. Because users lack understanding of insurance products and are not clear about their own positioning, even if users spend a long time, they may not necessarily find an insurance product that suits them. [0003] The insurance product recommendation technology currently used in the market recommends various insurance products for users. Although it has a certain orientation and reduces the user's query time, the insurance recommendation technology can only make recommendations based on the purchase ...

AI Technical Summary

Problems solved by technology

[0007] This application provides an insurance product recommendation method and system to solve the problem in the prior art that the insurance recommendation technology can only recommend a single recommendation, and the strength of the recommendation is not accurate enough

Images