Comment quality classification method, device, equuipment and readable medium

Abstract

The invention discloses a comment quality classification method, which comprises the following steps: obtaining comment data, and obtaining a quality evaluation reference index according to metadata feature analysis of the comment data; constructing a quality evaluation index system according to the language features of the comment data; constructing a classification model based on the quality evaluation index system, and classifying the comment data through a least square support vector machine of a recurrent neural network based on the classification model; and evaluating the classificationresult based on the classification model based on the quality evaluation reference index. The invention further discloses a comment quality classification device, computer equipment and a readable storage medium. According to the invention, a user comment quality evaluation system based on the FPGA is constructed, and when the data volume is gradually increased, high performance and processing rate can be kept; for a non-commercial platform, the comment quality is relatively high and more reasonable.

Description

technical field [0001] The present invention relates to the technical field of review quality evaluation, in particular to a method, device, device and readable medium for classifying review quality. Background technique [0002] With the rapid development of e-commerce, commercial shopping platforms have begun to collect user comment data, and display it to customers after sorting, categorizing, and importance ratings, so that they can quickly find satisfaction in the massive comments. Comments on demand. By optimizing the display method of comments, the user experience is improved, and the purchase rate of users is increased at the same time. [0003] Most of the existing review quality evaluation technologies are pure software implementations, which are transformed into classification and sorting problems. By extracting feature indicators related to reviews from different levels, supervised machine learning methods are used to classify and rank their quality. It is pred...

AI Technical Summary

Problems solved by technology

[0003] Most of the existing review quality evaluation technologies are pure software implementations, which are transformed into classification and sorting problems. By extracting feature indicators related to reviews from different levels, supervised machine learning methods are used to classify and rank their quality. It is predicted that the rate is low, takes up a lot of CPU resources, and requires regular maintenance; and there are few studies on non-commercial platforms, and user reviews are usually more casual, which cannot effectively reflect the true value of the product; most evaluation criteria are only based on usefulness voting , relatively one-sided

Images