Commendation method of personalized resource information based on scene information

Abstract

The invention discloses a commendation method of personalized resource information based on scene information. The method comprises: pre-processing the web page of a synergic labeling system, storing the information of all labeling behaviors of a user according to the information of the labeling behaviors, wherein the information is extruded by a specific user and comprises labeled resource information, used tag information and labeled time information; generating grading data expressing user preferences according to the tag information that the user uses the resources in a database and the time information of the labeled resource; calculating the similarity among users based on the generated grading data expressing the user preference so as to determine that the users with similar interest are adjacent; commending the resource that is not labeled by the user to the users according to the adjacent preference information of the users to complete the commendation of synergic filtrated personalized resource. Experiments show that better personalized commendation service can be provided for the users through integrating the scene information.

Description

technical field [0001] The invention relates to the technical fields of information filtering and data mining, and is a novel method for recommending personalized resource information based on situational information. Background technique [0002] With the development of Web 2.0, the amount of information on the Internet has increased exponentially. Massive information makes it difficult for users to quickly and conveniently discover and obtain the information they need. As an important means of information filtering, the recommendation system can automatically discover the information that users are interested in, and effectively provide users with personalized services. At present, the recommendation system has been integrated into major e-commerce systems, such as Amazon, eBay, Youtube, etc., and is still developing rapidly. DVD online rental company Netflix launched a competition on October 2, 2006: Netflix Prize, any organization or individual can get one million doll...

AI Technical Summary

Problems solved by technology

Due to the redundancy and ambiguity of labels, such as synonyms and polysemy, the model does not consider the natural language understanding of labels

The second model is to use tag information in the process of calculating resource recommendations. The shortcoming of this method is that it is difficult for the system to make recommendations when the overlapping usage rate of tags by users is very low.

This method defaults that the user has the same preference for all the resources he has purchased or browsed, and as time goes by, the user's preference remains static, so it cannot accurately describe the user's preference

Images