Paper Contents
Abstract
The increasing number of documents in unorganised textual form is making it harder to generate useful insights for analysis and decision-making. The research uses a combination of NLP techniques to automatically identify and extract both named entities and their related properties and values from free-text. It is typical for traditional NER to encounter problems with domain adaption, resolving context and getting related measurements or indicator data. Our solution involves running rule-based methods along with transformer-models for NER model, specifically BERT, RoBERTa and XLM-RoBERTa. We combine this with spaCy for NLP processing. To connect pronouns and entity references accurately and to precisely find EAV triplets in sentences, our approach includes special context tracking.When tShe data has been extracted, it is shown with statistical summary tables, as well as pie charts and bar graphs, to help understand the distribution of attributes among various entities. By testing on various input text, the system successfully showed it can be used for both kinds of metrics. The outcomes illustrate that applying NER in layers gives better context and improves performance. Because it is relevant to automated reporting, smart systems and healthcare, the proposed pipeline expands information extraction by creating a simple and transparent process for transforming text into data that can be analysed
Copyright
Copyright © 2025 DEEPAK KUMAR. This is an open access article distributed under the Creative Commons Attribution License.
