MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641126263 A) filed by Karpagam Academy Of Higher Education; Karpagam College Of Engineering; Dr. M. M. Karthikeyan; and Thameem Firthoush M on October 23, 2026, for Fake Product Review Monitoring And Removal For Genuine Rating.
Inventors include Dr. M. M. Karthikeyan; and Thameem Firthoush M.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: The exponential growth of e-commerce platforms has significantly increased consumer reliance on online product reviews and ratings for informed purchasing decisions. While reviews serve as a valuable source of feedback, the proliferation of fake and manipulated product reviews has emerged as a serious challenge, undermining consumer trust and distorting market fairness. Fraudulent reviews generated by bots, paid individuals, or coordinated groups often mislead customers by artificially inflating or deflating product ratings. Addressing this issue requires an intelligent, scalable, and automated solution capable of distinguishing genuine reviews from deceptive ones in real time. This invention presents an advanced system for monitoring, detecting, and removing fake product reviews to ensure genuine and trustworthy ratings across e-commerce platforms. The proposed solution leverages artificial intelligence, machine learning, and natural language processing techniques to analyze both the textual content of reviews and the behavioral patterns of reviewers. By integrating multiple dimensions of analysis, the system effectively identifies sophisticated fraudulent review activities that conventional rule-based or manual moderation systems fail to detect. The system performs in-depth textual analysis using natural language processing to evaluate sentiment polarity, semantic relevance, writing style consistency, and repetition patterns within reviews. These features help uncover subtle anomalies indicative of fabricated or manipulated content. In parallel, behavioral analysis modules assess reviewer activities such as review frequency, rating distributions, timing patterns, purchase verification status, and account usage behavior. Temporal analysis further strengthens detection by identifying unusual review bursts and coordinated posting activities. A unique credibility scoring mechanism is introduced to assign trust scores to individual reviews and reviewers. This score is computed by combining weighted outputs from content analysis, behavioral evaluation, and historical performance metrics. Reviews exceeding a predefined suspicion threshold are automatically flagged for removal or moderation, while genuine reviews are preserved for rating computation. This approach minimizes false positives and ensures fair treatment of legitimate user feedback. The system is designed for real-time operation and scalable deployment, making it suitable for high-volume e-commerce environments. Machine learning models are continuously trained and updated using newly labeled review data, enabling the system to adapt to evolving fraud strategies. An administrative dashboard provides platform operators with actionable insights, including flagged review statistics, reviewer credibility trends, and system performance reports. By ensuring that product ratings are calculated solely from verified and credible reviews, the proposed invention restores transparency and trust in online review ecosystems. The system not only protects consumers from misleading information but also supports honest sellers by preventing unfair reputation manipulation. Overall, this invention delivers a robust, intelligent, and adaptive framework for maintaining the integrity of product reviews and promoting confidence in digital marketplaces.
Disclaimer: Curated by HT Syndication.
