MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202641048584 A) filed by S. Hrushikesava Raju; Sri Venkateswara University; Subbaiahgari Ramachandraiah Ajitha; and Dr. Gajjala Venkata Ramesh Babu, Mangalagiri, Andhra Pradesh, on April 16, for 'a machine learning-based intrusion detection framework for classifying network flows into attack and non-attack categories.'
Inventor(s) include Subbaiahgari Ramachandraiah Ajitha; and Dr. Gajjala Venkata Ramesh Babu.
The application for the patent was published on May 1, under issue no. 18/2026.
According to the abstract released by the Intellectual Property India: "The present invention relates to a machine learning-based intrusion detection framework for classifying network flow data into attack and non-attack categories. The framework receives a flow dataset as input and performs preprocessing including data processing, missing value handling, duplicate value removal, data splitting, and label encoding to prepare the dataset for model training. An enhanced XGBoost algorithm is employed to train the classification model for identifying malicious and benign traffic patterns. The trained model is stored in a serialized format and deployed for prediction through a web-based framework. During operation, incoming flow data is analyzed by the trained model to determine whether the data corresponds to an attack or a non-attack. If the data is classified as an attack, the corresponding flow is ignored or rejected. If the data is classified as non-attack, the corresponding data is stored in a blockchain to ensure integrity, transparency, and reliability of the stored records. Thus, the proposed framework combines machine learning-based intrusion detection with blockchain-based secure data storage for improved network security and trustworthy record management."
Disclaimer: Curated by HT Syndication.
