MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202641074528 A) filed by Akash S; Ifthikhar Ali Khan; and Kavitha G on June 16, 2026, for A Code-Free Web Application For End-To-End Data Analytics With Near- Native Performance With Real World Datest.
Inventors include Akash S; Ifthikhar Ali Khan; and Kavitha G.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: A system and method for an integrated data analysis platform are disclosed. The Data Analysis Platform provides a unified environment for data ingestion, preprocessing, visualization, and machine learning. A preprocessing engine performs automated data cleaning, transformation, encoding, and feature engineering through structured and natural language commands. The system further includes a visualization module for interactive analytics and a machine learning module supporting multiple regression, classification, and clustering models with automated training and evaluation, all within a browser-based interface. The system further demonstrates improved predictive accuracy and optimized model performance through dynamic preprocessing, automated hyperparameter tuning, and model selection. Each machine learning model undergoes systematic hyperparameter optimization to enhance performance metrics such as MAE, RMSE, and R2 while minimizing error rates and training time. The platform integrates automated machine learning capabilities that generate a comparative performance leaderboard, ranking models based on evaluation metrics and execution time. Experimental results indicate that advanced ensemble models achieve near-optimal accuracy (R2 approaching 0.99-1.00) with significantly reduced error margins and efficient training times. The inclusion of automated tuning and leaderboard based model comparison ensures selection of the most accurate and efficient model, thereby improving overall analytical reliability, reducing time- to-insight, and enabling consistent high-quality predictions across diverse dataset
Disclaimer: Curated by HT Syndication.
