Spam Audit Detection Through Content Analysis 
Uzma A Mujawar1, Mallikarjun M Math2
1Uzma A Mujawar*, Computer Science and Engineering, KLS Gogte Institute of Technology,Belagavi, India.
2Dr. Mallikarjun M Math, Computer Science and Engineering, KLS GogteInstituteofTechnology,Belagavi, India.
Manuscript received on June 01, 2020. | Revised Manuscript received on June 08, 2020. | Manuscript published on June 30, 2020. | PP: 902-907 | Volume-9 Issue-5, June 2020. | Retrieval Number: E9895069520/2020©BEIESP | DOI: 10.35940/ijeat.E9895.069520
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Nowadays e-commerce has gained recognition in day to day life; hence network became tremendous source for gathering customer reviews/opinions by marketplace analyzers. The count of user reviews that merchandise receives is increasing at high velocity. Opinions being posted on social media differ significantly in superiority. The client needs to essentially go through all reviews regardless of their superiority and decide whether to purchase or not purchase the manufactured goods. The main difficulty in obtainable study on opinion assessment is that all opinions are considered regardless of the implication of each of them. Hence categorization of opinions depending on implication is an essential job. In this article, attempt is made for opinion assessment depending on its superiority, and help buyer make a proper buying judgment. A web mining technique that is narrative and effective is used to assess the consumer opinion for manufactured goods depending on marked allocation are anticipated. The superiority consideration of consumer reviews are classified as genuine, near duplicate, and duplicate opinion. It is carried out in three steps: (1) Recognize opinion regions to take out opinions. (2) Take out and separate features of reviews by quartile compute and assign weights to the features that belong to each group. (3) Consider the feature weights and group belongingness to assess the reviews. Investigational output demonstrates the usefulness of the proposed method which measures the quality of review and assesses it in view of that. The efficiency of client opinion summarization task is probably improved by recognizing and discarding irrelevant opinions.
Keywords: Review spam detection, feature extraction, feature comparison.