Financial management recommendation method, device and equipment based on behavior data

Abstract

The embodiment of the invention discloses a financial management recommendation method, device and equipment based on behavior data, and the method comprises the following steps: obtaining multi-dimensional attribute information and historical behavior data, wherein the multi-dimensional attribute information comprises financial management product multi-dimensional attribute information and corresponding user multi-dimensional attribute information; preprocessing the multi-dimensional attribute information and the historical behavior data, wherein the preprocessing comprises one or more of screening, clearness, missing value processing and singular value processing; inputting the preprocessed multi-dimensional attribute information into the constructed reinforcement learning model networkfor training to obtain recommended knowledge; and recommending the financing product to the target user according to the recommendation knowledge. By adopting the financial management recommendation method and device, the historical browsing behavior sequence information of the user is captured by using the reinforcement learning model, so that the financial management recommendation result is more accurate, and the click rate and the purchase rate of the user are greatly improved.

Description

technical field [0001] The present invention relates to the technical field of financial management intelligent recommendation, in particular to a behavioral data-based financial management recommendation method, device and equipment. Background technique [0002] With the deepening of inclusive finance, the intelligent wealth management recommendation market is becoming more and more mature. Not only are there a huge number of wealth management users, but their behavioral characteristics and preferences for wealth management products are also rich and diverse. Therefore, it is necessary for the recommendation system to make a targeted product ranking recommendation strategy for users with different characteristics, and to drive and guide the increase in the purchase rate of wealth management products. Most of the current recommendation systems are based on fixed rules, or based on the learning of product dimensions, or based on static indicators such as the similarity betwe...

AI Technical Summary

Problems solved by technology

[0003] 1. The final result in practice is that the purchase rate of recommended wealth management products is far from satisfactory

[0004] 2. It is impossible to use the dynamic information of the user's historical browsing behavior to portray the user portrait of the user

[0005] 3. User preferences will change over time. Traditional recommendation systems can only maximize current benefits, and cannot track and model dynamic changes in user interests and behaviors to obtain long-term benefits.

Images