Information recommendation method, information recommendation system and server

Abstract

The invention discloses an information recommendation system which comprises a business processing unit, a relation database and a KV database, wherein the business processing unit is used for sending history operation data of ht information to the relation database and the KV database in real time, reading the state data of the information from the relation database and the history operation data of the information from the KV database while receiving an information recommendation request, calculating a recommendation list of the information according to the history operation data of the information, the state data of the information and the user information, and returning the recommendation list; the relation database is used for calculating the state data according to the history operation data of the information; and the KV database is used for storing the history operation data of the information. The invention also discloses an information recommendation method. According to the scheme adopted by the invention, the information recommendation sensitivity is greatly improved, and the experience of the user is improved.

Description

Technical field [0001] The invention relates to Internet technology, in particular to an information recommendation method, system and server. Background technique [0002] With the development of Internet applications, the system can recommend information that may be of interest to the user, such as songs, videos, or products that the user may be interested in, based on the user's historical operation records. The current information recommendation service uses the method of offline calculation of the information recommendation list: the recommendation system calculates the information recommendation list that users may be interested in based on the user’s full historical operation records in the last three months at a fixed time every day. figure 1 It is a schematic diagram of the composition structure of the information recommendation system in the prior art, such as figure 1 As shown, the information recommendation system in the prior art includes a client 11, a business syste...

AI Technical Summary

Problems solved by technology

[0007] In the current information recommendation process, the business system 12 generally imports information historical operation data into the relational database 13 once a day. The reporting process generally takes 2 hours, and the relational database 13 calculates the information recommendation list that the user is interested in, generally takes 15 hours, and imports the information recommendation list into the business system 12, generally takes 3 hours, so, The process of recommending information may take more than one day, that is to say, the information that the user requests for recommendation on the day is likely to be the information that the user was interested in two days ago, which will reduce the sensitivity of the recommendation

If the relational database fails, it will also lead to failure to export the calculation results of a certain day, which will cause the recommendation results that users see today to be exactly the same as those seen yesterday, which will affect the user experience

And every time the recommendation algorithm is updated, the verification cycle of the effect will be very long

[0008] In addition, the current information recommendation list limits the number of information, and can only recommend a maximum of 100 information per day. If the user requests for recommendation too many times that day and all the information in the information recommendation list has been recommended, all the information will be re-recommended. The information in the recommendation list of the above-mentioned information causes the user to see repeated information, which affects the user experience

Images