Downloads: 145 | Monthly Hits: ⮙1

Review Papers | Computer Science & Engineering | India | Volume 7 Issue 3, March 2018

Fake Opinion and Brand Spam Detection Utilizing J48 Classifier Time Series

Sai Nawale | K. V. Reddy ^[2]

Abstract: As the innovation changes for reputation, approach to customary advertising additionally changes as individual to-individual correspondence to online surveys. As criticism, these online surveys are vital so client and to organizations or merchants. These surveys are useful for settling on choices in regard to nature of items and administrations. Organizations and merchants utilize opinions for settling on a choice for advertising procedures, execution to administrations or item, for development. Notwithstanding, the goals to all clients of clients are not valid for composing audits. These ideas, changes the substance of publicizing to customary, individual-to-individual correspondence to online reviews. These online reviews are vital to customer and to associations or dealers. In this paper, we proposed the strategy to perceiving the untruthful audits that are given by the clients which is having unmistakable semantic substance in light of slant examination as the surveys of films. In this paper creator speak to distinguish the spam untruthful surveys of motion pictures. For this arrangement, we utilized J48 classifier. We produce ARFF from the unmistakable elements to recognizing the untruthful audits. Utilizing Support Count as a part of Association Rules we additionally recognize Brands in Fake Reviews. The purpose of this paper is to propose a vivacious review spam detection system wherein the rating deviation, content based components and liveliness of reporters are used successfully. To beat the a fore said drawbacks, each one of these components are misleadingly inquired about in suspicious time breaks got from time course of action of overviews by a case affirmation framework. The proposed system could be an inconceivable asset in online spam filtering structures and could be used as a piece of data mining and learning disclosure assignments as a standalone system to channel thing review datasets. These systems can get compensate from our methodology to the extent time profitability and high accuracy.

Keywords: Brand Spam detection, Review Spam detection, ARFF, J48 classifier

Edition: Volume 7 Issue 3, March 2018,

Pages: 1019 - 1026

Similar Articles with Keyword 'J48 classifier'