As discussed supervised methods are class of algorithm that learn a model by looking at annotated training examples. Among the supervised learning algorithms for NER, considerable work has been done using Decision Trees[15], Hidden Markov Model (HMM)[19], Maximum Entropy Models (MaxEnt)[7], Support Vector Machines (SVM)[18] and Conditional Random Fields(CRF)[2]. Typically, supervised methods either learn disambiguation rules based on discriminative features or try to learn the parameter of assumed distribution that maximizes the likelihood of training data.