Optimized Association Word Knowledge Base Construction Using Apriori-genetic Algorithm

초록

The established simple mining technique extracts the association words considered semantic relation among words, it shows higher recall but lower accuracy of word clustering. AWKB(Association Word Knowledge Base) composed of association words saves many unsuitable association words in a class. In this paper, we propose word clustering for constructing OAWKB(Optimized Association Word Knowledge Base). For the purpose, we categorize web documents on the field of computer into 8 classes, and mine association words by Apriori algorithm. Then, we prune unsuitable association words in a class by Genetic agorithm. Finally, we can construct OAWKB. For the purpose of evaluating the performance of word clustering using Apriori-Genetic algorithm designed in this paper, we compare with the established Apriori algorithm and neural network algorithm.