{"title":"Summarization and Prioritization of Amazon Reviews based on multi-level credibility attributes","authors":"N. Aishwarya, Bhuvana L S, K. N.","doi":"10.1109/RTEICT52294.2021.9573744","DOIUrl":null,"url":null,"abstract":"It is common for most people to check Amazon reviews of competing products before a purchase decision. However, the ratings and reviews could be from a customer who has not actually purchased the product on Amazon. There is also a chance that the reviews are doctored by paid reviewers either to enhance a particular product's appeal or lower that of a competitor's product. So, the ratings and reviews do not always reflect the reality. We provide a mechanism to eliminate un-authenticated reviews prioritizing them based on their credibility score and summarize both positive and negative keywords. In this paper, we present a methodology to eliminate un-authenticated reviews based on three levels of pruning. The final rating of the review is adjusted using helpfulness ratio of the review, how old the review is and the helpfulness and experience of the reviewer. We also summarize the positive and negative keywords using ‘term frequency-inverse document frequency’ and ‘Long Short-Term Memory networks’. The reviews are prioritized the based on their credibility score. We were able to summarize and prioritize the reviews for easy analysis by the user. We also did an empirical validation of our proposed solution and found that the overall helpfulness factors improved.","PeriodicalId":191410,"journal":{"name":"2021 International Conference on Recent Trends on Electronics, Information, Communication & Technology (RTEICT)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 International Conference on Recent Trends on Electronics, Information, Communication & Technology (RTEICT)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/RTEICT52294.2021.9573744","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2
Abstract
It is common for most people to check Amazon reviews of competing products before a purchase decision. However, the ratings and reviews could be from a customer who has not actually purchased the product on Amazon. There is also a chance that the reviews are doctored by paid reviewers either to enhance a particular product's appeal or lower that of a competitor's product. So, the ratings and reviews do not always reflect the reality. We provide a mechanism to eliminate un-authenticated reviews prioritizing them based on their credibility score and summarize both positive and negative keywords. In this paper, we present a methodology to eliminate un-authenticated reviews based on three levels of pruning. The final rating of the review is adjusted using helpfulness ratio of the review, how old the review is and the helpfulness and experience of the reviewer. We also summarize the positive and negative keywords using ‘term frequency-inverse document frequency’ and ‘Long Short-Term Memory networks’. The reviews are prioritized the based on their credibility score. We were able to summarize and prioritize the reviews for easy analysis by the user. We also did an empirical validation of our proposed solution and found that the overall helpfulness factors improved.