Big data recommendation method, system and device for insurance industry

Abstract

The invention provides a big data recommendation method, system and device for the insurance industry, and the method comprises the following steps: (A) constructing an insurance industry big data center, storing the related data of a user in an HDFS system, and carrying out the distributed storage; (B) executing distributed data processing and computing through a Spark big data parallel computingframework, and processing log messages through a Kafka distributed message publishing and subscribing system channel; (C) dividing the users into different groups according to the attributes of the users based on an improved clustering model of demographic data, and executing a recommendation algorithm in different groups; and (D) executing front-end dynamic visual display through the ECharts open source visual library and the WebSocket full duplex communication protocol. The method can assist the insurance industry in commercial decision making, provides an insurance industry recommendationmodel, and guarantees the data safety.

Description

technical field [0001] The invention relates to the field of big data analysis, in particular to a big data recommendation method, system and equipment for the insurance industry. Background technique [0002] At present, the recommendation system has been widely used in many fields, such as news recommendation, Weibo recommendation, book recommendation, movie recommendation, product recommendation, music recommendation, restaurant recommendation, video recommendation, and the application scenarios of the recommendation system are becoming more and more diverse. It not only faces traditional problems such as data sparseness and cold start, but also faces more and more complex practical problems caused by big data. [0003] Such as figure 1 As shown, the existing recommendation system mainly has the following steps: [0004] 1. Data storage, storing the data to be trained by the recommendation system [0005] 2. Transfer the data to the computing computing center for prepr...

AI Technical Summary

Problems solved by technology

[0007]Because the quality of the recommendation system is closely related to the model training data, due to the protection of user privacy in the insurance industry, it is impossible to obtain a large number of actual users generated data, and cannot be artificially synthesized

Therefore, there are few recommendation systems for the insurance industry in the current market.

In addition, the insurance industry has high requirements for data security, so the recommendation system interface needs to be specially processed. Currently, the recommendation systems on the market have not done corresponding processing.

Images