The Improved Ontology kNN Algorithm and its Application

With the advances of the Web, more and more people, especially business people, use emails to communicate with each other. Hence, how to deal with business emails is becoming more and more important for decision makers, because among these emails, there hides valuable information such as the customer's complaints about a product or the interests of customers to a product. These are important information for a manager to propose marketing policies. In this paper, we develop an improved kNN algorithm----fkNN (fuzz kNN) algorithm based on ontology ideology to classify the emails. After classifying the emails into different classes, we can mine knowledge more easily based on the classified emails. Therefore, the classification effect is very important for mining knowledge further. Fortunately, our improved algorithm behaves much better than other algorithms in classification performance for our email datasets and other datasets.