An ensemble classification model for fake feedback detection using proposed labeled CloudArmor dataset. (July 2021)

• Record Type: Journal Article
• Authors: Taneja, Harsh; Kaur, Supreet
• Is Part Of: Computers & electrical engineering. Volume 93(2021)
• Publication Date: 2021-07
• Subjects: Cloud Service Provider -- Ensemble classification model -- Labeled CloudArmor dataset -- Fake Feedback detection -- Supervised Classification models -- Cloud Customer reviews; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2021.107217
• Languages: English
• ISSNs: 0045-7906

Abstract

Highlights: Ensemble model has been proposed to classify reviews for Cloud Service provider (CSP) posted by a cloud customer (CC) as fake or genuine on labeled CloudArmour dataset. Labeling is achieved through: Number of Reviews by a CC to single CSP; Number of CSPs reviewed by a CC; & Frequency of feedbacks published. Proposed ensemble model is compared with various classifiers. Ensemble model has depicted highest values as: 97.51%; 98.19%; 93.65%; & 95.86% respectively in terms of accuracy, precision, recall and F1-Score. Abstract: With an increase in the use of Cloud Services, Cloud Customers (CCs) have started browsing reviews about various Cloud Service Providers (CSPs) online provided by numerous CCs to gather their sentiments. This pattern has proved to be fruitful, but it can become ineffective if fake reviews are posted. This article aims at introducing an ensemble model to classify reviews as fake or genuine on the proposed labeled CloudArmour dataset. Labeling has been achieved through these aspects: Number of Reviews provided by a CC to single CSP; Number of CSPs reviewed by a CC; & Frequency of feedbacks published. Additionally, we have compared the performance of the proposed ensemble model with these supervised machine learning classification techniques: KNN; Logistic_Regression; and SVM_rbf on the basis of the following parameters: accuracy; precision, recall; and F1-Score. Ensemble model has yielded highest values as: 97.51%; 98.19%; 93.65%; & 95.86% respectively. Thus, we conclude that the ensemble model has outperformed other models. Graphical abstract: Image, graphical abstract