Protecting Against Malicious Code Injection in Reviews on Web Applications

Abstract

Malicious code injection is done by the attackers or hackers mainly in reviews, these are the fake user identities that are created by the attackers, through sending continuous links to some user till the user clicks on that link. When user clicks the link, that particular users identity will be stolen by the attackers which is done like the phishing, generally these reviews which are maliciously injected are mixed with the original users review of the product in a website. To identify these malicious injection attacked reviews, in this paper the Naive Bayes Classifier (NBC) algorithm is used. Then to eliminate the unwanted disturbed data in the website the Natural Language Processing technique is used. The Natural Language Processing (NLP) technique, which removes unwanted data by understanding the text or words in review. Then the length of the reviews in the given sample dataset is reduced for easy understanding by using the Principal Component Analysis (PCA) algorithm which reduces the dimensionality of the reviews. Then user input review is compared with the sample dataset and classified as good review and bad review and also detected as malicious or not by using the Naïve Bayes Classifier algorithm which is used for the classification of the objects.