A Hybrid Method for Automotive Entity Recognition Tata Elxsi, Limited

Format:

Book

Conference/Event

Author/Creator:

Sivaraman, Navya K., author.

Contributor:

Koduri, Rajesh

Manalikandy, Mithun

Conference Name:

SAE WCX Digital Summit (2021-04-13 : Live Online, Pennsylvania, United States)

Language:

English

Physical Description:

1 online resource cm

Place of Publication:

Warrendale, PA SAE International 2021

Summary:

Over the past decades, automotive industry has made substantial investments in automation solutions, electric and autonomous vehicles and advanced product technologies for enhancing vehicular communication and so on. The rise of industry 4.0 brings out a revolutionary transformation in the automotive industry with low-cost computing, high-speed connectivity, and machine learning that have enabled the digitization of the physical world, transforming insights into optimized actions. These technologies have an important role in the growth and future of automotive domain. Hence it is relevant and important to get insight of different automotive entities like OEMs (Original Equipment Manufacturer), geographical locations and their focusing technologies adapted currently and in the nearby future. In this paper, we are implementing a hybrid method for entity recognition, which is a combination of both rule-based and machine learning based entity recognition techniques. Rule based entity recognition is achieved by executing a certain series of rules on raw data by using multiple resource like lexical resources, customized gazetteers and multiple pattern bases. Machine Learning based entity recognition is done based on Hidden Markov Model (HMM) machine learning model, which express long-distance-dependent and overlapping features. In this paper, we annotate our own training and testing data sets to use in the related phases for identifying entities. Finally, we are combining the results of both rule based and machine learning based methods. Our approach to entity recognition using hybrid method achieves 91.2 percent accuracy

Notes:

Vendor supplied data

Publisher Number:

2021-01-0179

Access Restriction:

Restricted for use by site license