Identifying The Established Semantic Kinds In Text Using Named Entity Recognition

Named entity recognition (NER) is a natural language processing (NLP) method that extracts information from text. NER includes identifying and ordering significant data in text known as named elements. By approaching span-based method (where tokens are represented as spans) we can detect named entity for each span. However some current algorithms check every span separately without knowing previous span, which leads to ambiguity. To overcome this, a bi-directional two-dimensional operation is used, which uses knowledge of previous spans and compares the two spans and generates meaningful named entities. For implementation of the NER we have used the three public datasets. The experimental results in this paper shows that the method we used is very effective to generate meaningful nested named entities.