Internet forum user interest modeling method based on uncertainty treatment

Abstract

The invention belongs to the technical field of internet user behavioral analysis, in particular to an internet forum user modeling method based on certainty, which comprises the steps of: introducing a subordinate function for expressing user interests during the modeling, computing parameters of the function by adopting a similar Gaussian type subordinate function based on behavioral characteristics of a user in the forum; extracting a user interest text by adopting a text processing method, computing user interest text vectors according to different weight configurations; establishing a user interest model in a higher-dimensional space formed by the text vectors and a subordinate function discourse; and describing interest distribution of the user in different topic spaces by adopting a probability density function. The model established by the invention can reflect the vagueness of the user interests and express the randomness of the user interests, and enable the expression of the user interests to more approach the user requirement, thereby being more reasonable and being used for various analysis occasions based on internet user interests.

Description

technical field [0001] The invention belongs to the technical field of network user behavior analysis, and in particular relates to a modeling method oriented to user interests of network forums. Background technique [0002] With the rapid promotion of the application of Web2.0 on the Internet, many interactive forum websites have appeared. These websites gather a large number of Internet users, where they publish posts, reply to posts, and show different interests in posts on different topics. For many commercial applications, accurately discovering user interests and discovering more interest groups is the primary condition for successful business development. Therefore, it is an effective way for commercial applications to acquire user groups by making full use of the behaviors of users posting or replying to posts in these forum websites to tap users' interests. [0003] At present, people's research on the interests of Internet users mainly focuses on the user's sear...

AI Technical Summary

Problems solved by technology

[0004] 1. Describing user interests purely in a probabilistic manner can only explain the randomness of users’ interest in a certain topic, but not the ambiguity of this interest, and ambiguity is an important aspect for people to analyze and understand user interests

[0005] 2. There are great differences in the number of times and lengths of postings or replies by users of online forums, which reflects the differences in user interests to a certain extent. However, the existing models only express relevant texts or click behaviors of users, which inevitably These important features that reflect user interests are lost

[0006] It can be seen that improving the randomness and ambiguity description ability of user interests is very important to describe user interests more reasonably, and making full use of user behavior characteristics on forums is critical to accurately express user interests, while existing methods in These two aspects are still unable to meet the requirements of user interest modeling analysis

Images