Statistical Dictionary with Conditional Random Fields to Identify the Kannada Named Entities
M. Pushpalatha1, Antony Selvadoss Thanamani2

1M. Pushpalatha, Maharani’s Science College for Women, Mysuru (Karnataka), India.
2Dr. Antony Selvadoss Thanamani, HOD, Department of Computer Science, NGM College of Arts and Science, Pollachi, Bharathiar University, Coimbatore (Tamil Nadu), India.
Manuscript received on 24 April 2019 | Revised Manuscript received on 02 May 2019 | Manuscript Published on 08 May 2019 | PP: 388-392 | Volume-7 Issue-5S3 February 2019 | Retrieval Number: E11700275S19/19©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (

Abstract: We present an algorithm to recognize and identify the named entities of Kannada text document. The Kannada text document is collected from Central Institute of Indian Languages has many issues to be addressed. The proposed method has addressed the objective of algorithm is to determine and recognize the Kannada Named Entities like name of a person, designation of a person and place needs to be identified and recognized. The proposed statistical dictionary with conditional random fields in deep neural networks have been used to achieve the task of recognition of Kannada Named Entities The dictionary of Kannada words is formed from the statistical approach of matching patterns of Unicode values of individual words of a document. The sequence of Unicode values are considered for matching of patterns with deep architecture of neural networks has helped us in recognizing the Kannada word items from a dictionary formed from the proposed method. Finally the proposed method has achieved an accuracy of 84.46% from the proposed statistical dictionary of Kannada words with Conditional Random Fields.
Keywords: CRF, Dictionary, Deep Learning, KNER.
Scope of the Article: Deep Learning