Detecting and Extracting Named Entities with Particular Reference to Marathi Language

Abstract

The organization name, person name, location name, brand name and others are called named entities. The purpose of detecting and extracting named entities is to recognize all the named entities in the document and extracting those named entities. Detection of named entities is two step procedure- proper nouns identification and the classification of identified proper nouns. In the first step proper nouns are recognized from the text. In the second step proper nouns are classified into the different classes like the names of an organization, person, location, brand and others. Recognition of named entities is used in many applications like Natural Language Processing, Machine Translation and Machine Learning. Morphologically rich and free ordered features are present in Indian languages. Reorganization of named entities is difficult in the Indian languages like Marathi, Hindi, Urdu, Telugu and Bengali etc. The objective of this paper is to conduct the survey on recognition of named entities in different Indian languages and compared the performance metrics of different named entity approaches. Also, mentioned the challenges of Named Entity Recognition in Marathi language like morphological features, no capitalization, writing variations and ambiguity.