An Empirical Study on Spectral Clustering-based Software Defect Detection

Qing, Mingshuang; Ge, Xiuting; Hui, Zhanwei; Pan, Ya; Yi, Fan; Wang, Xiaojuan; Cao, Xu

doi:10.1109/dsa52907.2021.00012

Cited by 8 publications

(1 citation statement)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In terms of false comment group detection, a spectral clustering group detection algorithm based on commenter similarity matrix is proposed by Ye Zicheng et al [5] , and Zhang Qi et al propose a sailor group detection algorithm through the selection of features of sailor group counterfeiting behaviors, the construction of comment graphs with weights, the screening of suspicious subgraphs, and clustering of sailor groups based on community discovery algorithms [6] . For the research on false comment group detection, most scholars use frequent item mining and clustering algorithms to obtain candidate groups and manually label them.…”

Section: Introductionmentioning

confidence: 99%

Research on the identification method of fake reviews in e-commerce

2024

AJBM

View full text Add to dashboard Cite

With the development of e-commerce, many merchants, to lure consumers to buy their own goods, will create the illusion of hot stores and excellent goods by means of false reviews. To identify false reviews, this research paper proposes a false review identification method, which mainly uses entropy weight TOPSIS model, K-Means cluster analysis and logistic regression. The article selects evaluation indexes such as sentiment polarity, text length, review usefulness and rating deviation degree, and calculates the review score by entropy weighting method of TOPSIS model. Subsequently, K-Means clustering was used to categorize the reviews into two groups: customer reviews and machine reviews and validated by logistic regression. The experimental results show that the method has a good ability to recognize false reviews with high accuracy, recall, precision, and AUC value. Taken together, the method provides an effective solution for false comment recognition with practical application potential.

show abstract

Section: Introductionmentioning

confidence: 99%