MUMBAI, India, Jan. 23 -- Intellectual Property India has published a patent application (202511127388 A) filed by Greater Noida Institute Of Technology; Dipali Singh; Beauty Jaysawal; Archana; Deepal Sharma; Arpita Gaur; Shazia Aiyub; Deepika Kumari; and Anupma Surya, Greater Noida, Uttar Pradesh, on Dec. 16, 2025, for 'fake account protection in social media using machine learning.'

Inventor(s) include Dipali Singh; Beauty Jaysawal; Archana; Deepal Sharma; Arpita Gaur; Shazia Aiyub; Deepika Kumari; and Anupma Surya.

The application for the patent was published on Jan. 23, under issue no. 04/2026.

According to the abstract released by the Intellectual Property India: "The present invention relates to an intelligent system for detecting and preventing fake accounts on social media platforms using machine learning algorithms. The system analyzes multiple features including profile completeness, posting frequency, friend network patterns, content behavior, follower ratios, and engagement metrics to distinguish authentic accounts from fraudulent ones. The invention employs a hybrid feature selection approach combining behavioral and structural attributes, processed through a custom preprocessing pipeline that handles missing values, removes noise, and balances datasets. Multiple classification algorithms including Logistic Regression, Decision Trees, Support Vector Machines, and Random Forest are implemented, with Random Forest demonstrating superior performance. The system provides real-time automated detection, eliminating manual verification needs while maintaining high accuracy and low false positive rates. An early warning mechanism flags suspicious accounts before they cause widespread harm, offering proactive protection against misinformation, phishing attacks, scams, and cyberbullying. The lightweight model design ensures efficient operation with limited computing resources, making it suitable for deployment across platforms of varying scales while significantly reducing moderation costs and enhancing user trust."

Disclaimer: Curated by HT Syndication.