Evaluate spam detection using hybrid technique of Support Vector Mechine

  Kishan Thumar,  Prof. Dulari Bosamiya

Due to the social media, online reviews are widely adopted. They greatly impact consumer’s purchasing decisions. A slight difference in a business rating on a review website can significantly change the reputation of them. Review spamming is quite common on many online shopping platforms like Amazon. In this work a hybrid method is used to detect whether a review is spam or not. From previous studies it is found that SVM(Support Vector Machine) gives comparatively good result for this task. PSO (Particle Swarm Optimization) is the optimization technique which can be used improve the result of SVM. Here duplicate detection is used after preprocessing of data to find if two reviews are same or not. After that feature identification is done and then PSO is used to choose the population and these values are used in SVM to classify reviews. Hence in this work, a Hybrid classifier (SVM+PSO) to classify review as a spam or not spam is used.

Keywords- Data mining, Web mining, Review Spam, Review Spammer, Support vector machine(SVM), Particle swarm optimization(PSO)

Unique Identification Number - IJEDR1504143
Page Number(s) - 807-811
Pubished in - Volume 3 | Issue 4 | December 2015
DOI (Digital Object Identifier) -   
Publisher - IJEDR (ISSN - 2321-9939)

  Kishan Thumar,  Prof. Dulari Bosamiya,   "Evaluate spam detection using hybrid technique of Support Vector Mechine", International Journal of Engineering Development and Research (IJEDR), ISSN:2321-9939, Volume.3, Issue 4, pp.807-811, December 2015, Available at :http://www.ijedr.org/papers/IJEDR1504143.pdf