Commodity comment data labeling system and method based on hierarchical AP clustering

Abstract

The invention provides a commodity comment data labeling system based on hierarchical AP clustering. The system includes a data capturing module, a word vector training module, a feature information extraction module and a feature information labeling module. The data capturing module stores corpus information and comment data. The word vector training module obtains a training corpus set. The feature information extraction module obtains a feature information set corresponding to the comment data. The feature information labeling module obtains the comment data labelled result after clustering. The beneficial effects of the invention are as follows: the commodity comment data labeling system and method based on hierarchical AP clustering are provided, the purpose of automatically labelingcomment data is achieved, the value orientation of the feature information can be mined and presented in the form of labels to merchants and customers, support is provided for subsequent data analysis, and companies and consumers are further provided with a convenient, scientific and intuitive tool for obtaining useful comment information.

Description

technical field [0001] The invention relates to the fields of computer data processing and data mining, in particular to a system and method for labeling commodity review data based on hierarchical AP clustering. Background technique [0002] In recent years, the vigorous development of e-commerce has led more and more people to choose the consumption mode of online shopping, and the transformation of consumption patterns has also brought new opportunities and challenges to enterprises and merchants. In the virtual environment of online shopping, a large number of research and survey results show that online reviews are the most important factor for consumers to make purchase decisions. At the same time, online reviews as feedback data can also help companies improve products, understand user needs, and enhance competition. strength and reputation. According to Nielsen’s survey data in the first half of 2014, 70%-80% of consumers will check product reviews as important refe...

AI Technical Summary

Problems solved by technology

[0004] However, with the advent of the era of big data, the amount of review data has increased rapidly, and the review data itself has characteristics such as irregularity and redundancy, making it difficult to obtain useful review information.

The number of comments on e-commerce websites is growing exponentially, and these comments have different styles, and there is no standard model for language structure to follow... Faced with a huge number of comments, it is difficult to mine user comments only by manual analysis and processing. time consuming and inefficient

[0005] In major e-commerce websites such as JD.com, Taobao, and Yihaodian, some user reviews have been categorized to provide a more intuitive and convenient experience for customer inquiries through label classification, but the disadvantage is that these labels Pre-set by merchants, most of them are positive labels, which cannot fully reflect customers' comments on products; in addition, labels are chosen by reviewers themselves, but it can be found through statistical data that only about 10% of reviewers are willing to put a label on the review. labels, and 90% of the review data has no label information, which makes it impossible to make full use of the evaluation orientation implied by each review

Overall, the comment labeling technology in practical applications is still in the artificial stage, and a large amount of valuable comment information has not been fully utilized

Images