Svm Based Improvement in Knn for Text Categorization

ABSTRACTIn today‟s library science, information and computer science, online text classification or text categorization is a huge complication. [1]With the enormous growth of online information and data, text categorization has become one of the crucial techniques for handling and standardizing text data. Various learning algorithms have been applied on text for categorization. On the basis of accuracy and efficiency KNN (K Nearest Neighbour) algorithm prove itself to be very efficient algorithm as compared to other learning algorithms. The framework of KNN with TF-IDF is studied and some changes need to be done for removing time complexity and improve accuracy so, proposed work is based on using SVM classifier which helps in splitting of training and testing data and take less time from the previous work with iKNN (improved KNN) algorithm which gives less time and more accuracy and overall improve text categorization.