Research Of Web Document Classification Based On Positive And Negative Association Rules

Association rule is an important part of data mining, is in order to find the links between items.The mining association rule technique is used in Web Document Classification, could more effectively organize and manage the vast amounts of Web information, and more quickly find the information on the network. However, the majority of researchers only use positive association rule in the Web document classification, and less involved negative association rules in it. Negative association rules could find the negative associations between itemsets, and also a complementary for positive association rule mining.Use negative associations in Web document classification can find the negative associations among Web documents, and also to enhance web document classification accuracy. How to apply negative association rules to Web document classification is a new problem, this paper discussed it, and presents a method of Web documents classification based on positive and negative association rules.This article describes the algorithm of Web document classification, and summarizes the current actuality of positive and negative association rules in domestic and overseas, and presents a method of Web documents classification based on positive and negative association rules. Firstly, Web documents is preprocessed to change the unstructured data into structured data, in order to establish a new set of items; then use Apriori algorithm to form the frequent 2-item sets,and use the modified PNARC model to select rules,and then removes contradictory association rules to get the right positive and negative association rules,so the relativity of document could be distinguished, and could determine that whether the documents are compartmentalized to the same class, finally through experiments to validate the method, show that the algorithm can improve the accuracy of Web document classification.