Collaborative filtering and recommendation method for real-time hot learning resources of primary and secondary school students

Abstract

The invention provides a collaborative filtering and recommendation method for real-time hot learning resources of primary and secondary school students. The method includes the following steps of: 1)firstly starting a zookeeper program and a kafka message queue cluster, and configuring the kafka message queue cluster in the zookeeper program; 2) submitting a user read-write request to a server terminal by a client, and firstly writing the user read-write request in the kafka message queue cluster after the server terminal receives the read-write request of the client; 3) taking out read-write request messages that need to be processed by a service module inside the server terminal; 4) performing noise treatment by the service module through a de-noising module; 5) submitting user requests and de-noised data to a Spark Streaming framework to perform real-time treatment; 6) writing ranking results in a redis module; and 7) displaying results after the client receives messages.

Description

technical field [0001] The invention relates to a collaborative filtering method, in particular to a method for real-time hotspot collaborative filtering and recommendation of primary and middle school students' learning resources based on the Spark framework. Background technique [0002] In the learning resource system for primary and middle school students, in addition to the search of learning resources through the question bank and the recommendation of learning resources based on user behavior data, students know which learning resources are the most popular at this moment. A review with the highest score is also a very important feature. In addition to knowing the learning resources recommended by students, if they can know that everyone is most concerned about, the most people participate in the discussion, and the most incisive learning resources are obtained through the most people's ratings, then students can go to participate in these learning resources together ...

AI Technical Summary

Problems solved by technology

In the existing technology, how to achieve the goal and make the function meet the expectation is designed. However, this is not feasible in practice. The reason is that the existing technology lacks data backup and data integration. Strategy, so that there will be no problems in normal times, and there will be no problems when everything comes as expected, but when encountering emergencies or emergencies, the entire system will be vulnerable and extremely fragile

[0005] 2. Lack of reliable message processing mechanism

All these behaviors will lead to confusion in system processing, and even make the entire system unavailable, which seriously damages the interests of users.

[0006] 3. Not optimized enough for real-time processing

In the existing technology, although real-time computing is performed based on spark, and the real-time performance of HBase is used as the data persistence method, this implementation method can have a lot of room for improvement and improvement.

Although HBase takes real-time as its major advantage, as long as the data exceeds a threshold, it will eventually be stored on the hard disk, so it will be slower than the read and write cache when reading and writing, which will affect real-time processing. The speed makes the real-time processing not meet the real-time requirements

Images