A Hybrid Statistical Approach for Named Entity Recognition for Malayalam Language
Jisha P. Jayan, R R Rajeev, Elizabeth Sherly · 2013
Named-Entity Recognition (NER) plays a significant role in classifying or locating atomic elements in text into predefined categories such as the name of persons, organizations, locations, expression of times, quantities, monetary values, temporal expressions and percentages. Several Statistical methods with supervised and unsupervised learning have applied English and some other Indian languages successfully. Malayalam has a distinct feature in nouns having no subject-verb agreement, which is of free order, makes the NER identification a complex process. In this paper, a hybrid approach combining rule based machine learning with statistical approach is proposed and implemented, which shows 73.42% accuracy.