Automatic pairing selection method and device in data processing system

Abstract

The invention discloses an automatic pair selection method and a device in a data processing system. The method comprises the steps: A, receiving a pairing request aiming at a user; B, detecting state information of the user at a current moment, selecting a mapping relation through a state in a preset reinforcement learning function and a type of a paired object, and calculating a selected probability of types of different paired objects which correspond to the state information of the user at the current moment; C, selecting one paired object according to the selected probability; D, detecting information of reaction motion of the user to the paired object, and carrying out feedback corrections on the corresponding states in the reinforcement learning function and the selecting probability of the types of the corresponding paired objects according to the obtained information of the reaction motion. The device comprises a request receiving module, a detection module of the state information, a reinforcement learning module, an object selection module, a detection module of the reaction information and a correction module. Through the automatic pair selection method and the device in the data processing system, a relevance degree between the selected paired object and a pairing request party is improved, and accordingly a final success rate of pairing is improved.

Description

technical field [0001] The invention relates to network data processing technology, in particular to an automatic pair selection method and device in a network data processing system. Background technique [0002] At present, with the development of Internet technology, network data processing systems in various subdivisions have emerged and developed one after another, meeting the special needs of users in different subdivisions. For example, the communication network system can meet the user's real-name or anonymous communication needs, the social network system can meet the user's social needs, the e-commerce platform system can meet the user's commodity purchase needs, and the online blog system can meet the user's log display needs. The system can meet the user's reading needs and so on. [0003] In a network data processing system, in many cases the system needs to perform selection and pairing processing according to the user's request. For example: select anonymous...

AI Technical Summary

Problems solved by technology

[0007] The disadvantage of this random pairing method is: the correlation between the randomly selected pairing partner and the pairing requester is extremely low, and users are often dissatisfied with the pairing partner randomly selected by the system, resulting in a very low final pairing success rate

[0010] 1) This method artificially designs a matching strategy based on the preferences of most users and acts on global users, while ignoring the individual needs of different users, resulting in low correlation between some user status information and paired objects

[0011] 2) The user's preferences for anonymous communication partners are likely to change with different dates (such as different working days or holidays) and different time periods. Using this artificial static pairing method will not be able to adapt to this dynamic user status Due to changing requirements, under certain dynamic conditions, the correlation between user status information and paired objects is not high

[0012] 3) Since this method is based on manual matching based on experience, the estimated value is generally used when setting the matching probability, so it is difficult to give a pairing object that is highly related to the recipient

[0013] 4) Slow response to user feedback

Although the manually set static pairing strategy can adjust the pairing strategy by observing and analyzing the user's usage over a period of time, this feedback mechanism has a long cycle and cannot quickly make strategic adjustments to the user's usage

[0014] In short, the existing network data processing system selects a pairing object for the pairing requester. The correlation between the selected pairing object and the state (including static state and dynamic state) of the pairing requester is not high. The pairing requester often Unsatisfied with the matching result, resulting in a low success rate of final matching

Images