NAMED ENTITY RECOGNITION (NER) FOR NEWS ARTICLES

Abstract

Named Entity Recognition (NER) plays a pivotal role in automating the extraction and categorization of named entities from textual data, enabling efficient information retrieval and analysis across various domains. This paper presents a comprehensive study on NER techniques, focusing particularly on their application in news articles. The project employs Conditional Random Fields (CRF) as a discriminative probabilistic model for sequence labeling tasks, leveraging feature engineering and preprocessing steps for accurate entity recognition. The CoNLL-2003 dataset serves as the benchmark dataset for training and evaluating the CRF model, showcasing its performance in identifying entities such as persons, organizations, and locations.