Extracting Information from Indian First Names

First name of a person can tell important demographic and cultural information about that person. This paper proposes statistical models for extracting vital information that is gender, religion and name validity from Indian first names. Statistical models combine some classical features like ngrams and Levenshtein distance along with some self observed features like vowel score and religion belief. Rigorous evaluation of models has been performed through several machine learning algorithms to compare the accuracy, FMeasure, Kappa Static and RMS error. Experimental results give promising and favorable results which indicate that these models proposed can be directly used in other information extraction systems.