MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641122376 A) filed by Dr. Theresa Cenate C. F; V Dharshan; Ss Rachna; and V Karthikeyan on October 13, 2026, for Qpredictai: Intelligent Question Paper Prediction.
Inventors include Dr. Theresa Cenate C. F; V Dharshan; Ss Rachna; and V Karthikeyan.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: ABSTRACT QPredictAI is a web application designed to address the problem of identifying similar examination questions from previous years' papers. The tool enables the users to upload the required PDF files through a web interface built using FastAPI. The extraction of information takes place with the help of PyMuPDF, an open-source Python library. The text is subjected to pre-processing, which includes converting text to lower case, removal of unwanted symbols and white spaces. The pre-processed text is divided into questions using Regular Expressions. The questions are then converted into an array of numbers using TF-IDF Vectorizer. With the utilization of Cosine Similarity to determine the similarities between the text strings, a threshold value assists in deciding the repeated questions. Finally, a dashboard is designed to display the results, and the questions can also be organized according to frequency. QPredictAI has been built using Python, FastAPI, PyMuPDF, and Scikit-learn. It is interesting to note that the tool can reduce the efforts required in comparing multiple documents for finding similar questions. Future scope includes the implementation of Optical Character Recognition (OCR) for scanned documents and the utilization of semantic search and language models for improving accuracy.
Disclaimer: Curated by HT Syndication.
