Fine granularity evaluation information mining method and system

Abstract

The invention discloses a fine granularity evaluation information mining method. The method includes the steps of adopting an initial evaluation object seed set and an initial evaluation word seed setas bases, utilizing a semi-supervised iteration method to extract a relevant structure of a candidate evaluation object set and a candidate evaluation word set, and separately adopting evaluation objects which are most relevant to candidate evaluation words and are mined from the candidate evaluation object set as new evaluation objects; in a similar way, adopting evaluation words which are mostrelevant to candidate evaluation objects and are mined from the candidate evaluation word set as new evaluation words, adding the mined new evaluation objects and the mined new evaluation words into the initial evaluation object seed set and the initial evaluation word seed set, iteratively and alternatively mining potential evaluation information, and extracting an evaluation object and an evaluation word which correspond to the evaluation information; additionally, carrying out visualization display on the evaluation information corresponding to evaluation sentences to provide convenience for a user to directly check. The invention also discloses a fine granularity evaluation information mining system.

Description

technical field [0001] The invention relates to the technical field of network data processing, in particular to a fine-grained evaluation information mining method and system. Background technique [0002] With the continuous popularization of social networks and mobile Internet, the cost for people to publish information is getting lower and lower, and more and more users are willing to share their opinions and comments on people, events, and products on the Internet, and Internet companies are becoming more and more It is hoped that automated tools and techniques can be used to obtain or mine the opinion information contained in these comment statements. Mining user opinion information refers to discovering content with the author’s subjective emotions from the text. User opinions usually include information such as evaluation objects, evaluation words, evaluation emotions, evaluation owners, and publication time. Among them, evaluation words are varied and updated quickl...

AI Technical Summary

Problems solved by technology

[0004] At present, researchers such as Bing Liu mainly perform word segmentation and tagging on comment sentences, and then use association rules to mine frequent items as candidate evaluation objects, and pruning them, but this method does not effectively screen candidate evaluation objects , it is necessary to further process all the extracted candidate evaluation objects, and the processing results cannot guarantee the recognition accuracy of the comment information

Images