Fake product Review Classification and Analysis using AI
Fake product Review – Abstract: The impact of reviews on any e-commerce site is of great importance, as it can be the basis for a buyer’s decision to buy any product. The buyer tries to evaluate the authenticity and quality of the product using the feedback given by other previous buyers in the form of a review. But, sellers are taking advantage of the reviews by posting reviews in an attempt to promote or defame a product. Such reviews which are not a genuine opinion of an individual are termed fake reviews. The existence of such fake reviews makes the buyer unable to make the right judgments of sellers, which can also cause the credibility of the platform to be downgraded. Thus, it is very important to identify the fake reviews on the platform. In this paper, we propose a method to detect such fake reviews using a logistic regression model by considering review-centric features achieving an overall accuracy of 82%. Also, our study illustrates the impact of the “verified purchase” feature in fake review classification.
Collecting all the Review data through the internet and just store into the database is implemented on Existing System.
They just collect the details only.
They should not implement the Prediction work.
In Our Prosed System, we implement the naive Bayes algorithm for prediction. Naive Bayes classify the past data and generate the prediction output. To design and build an efficient and robust fake review detecting model for e-commerce product reviews using a Naïve Bayes model which follows the idea of modeling the probability for classifying fake and genuine reviews. Also, provide a consensus strategy for feature extraction and text preprocessing.
The Advantage of the naive classifier is that is simple and converges quicker.
Easy to implement this system and get the prediction
- OS – Windows 7, 8, and 10 (32 and 64 bit)
- RAM – 4GB
In this paper, we have discussed the impact of the logistic regression model for identifying fake reviews using review-centric features. Along with review content, we have provided a set of review-centric features for the classification of fake reviews. One of the review-centric features we propose in this paper is “verified purchase”. Our research shows that using “verified purchase” as a feature for classifying fake reviews has an outstanding effect. In addition to this, we have proposed two feature extraction techniques viz. Tf-idf and Count Vectorizer and therefore, conclude that implementing logistic regression with Count Vectorizer on the used dataset has achieved an accuracy of 82% and an accuracy of 81% with Tf-idf.
 Barbosa, Luciano & Feng, Junlan. (2010). Robust Sentiment Detection on Twitter from Biased and Noisy Data. Cooling 2010 – 23rd International Conference on Computational Linguistics, Proceedings of the Conference. 2. 36-44.
 Enhanced Sentiment Learning Using Twitter Hashtags and Smileys Dmitry Davidov, Oren Tsur, ICNC / 2, Institute of Computer Science The Hebrew University 2010.
 Go, Alec & Bhayani, Richa & Huang, Lei. (2009). Twitter sentiment classification using distant supervision. Processing. 150.
 “Fake review detection using opinion mining” by Dhairya Patel, Aishwerya Kapoor and SameetSonawane, International Research journal of Engineering and technology (IRJET) , volume 5, issue 12,Dec 2018.