MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202541012448 A) filed by Vivek Joy on February 13, 2025, for Method And System For Conversion Of Unstructured And Semi-Structured, Tabular, Textual Data To Structured Data.
Inventor includes Vivek P Joy.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: A method and system are disclosed to create a structured dataset (108) from unstructured or semi-structured, tabular, textual data (101A – 101G). The method discloses the steps to input data could be from computer files or from paper. The method includes processor implemented steps to convert input data (101A – 101G) to a structured dataset where relationships between different columns are specified. The method detects relationships using three different approaches. These are Column Name Based Relationship Detection (501), Column Data Based Relationship Detection (502) and Machine Learning based Relationship Prediction (503). The relationships discovered by the different methods are aggregated and a confidence score is generated. This is compared with a threshold (507) where the threshold is configuration (103) parameter. The relationships meeting the threshold criterion can be optionally validated humans. The method incorporates the newly found relationships (508) into an output which is a structured dataset (108).
Disclaimer: Curated by HT Syndication.
