Research On Data Mining Technology And Its Application On Web Services

As a distributed computing technology used to build the Service-Oriented Architecture(SOA), web service has good interoperability, self-description and modular features. And web service has become a focus of attention for both industry and academia because of the advantages on loose coupling, standardization, and highly integrated capacity etc. However, with the great development of web service, the number of web services increases rapidly. There are so many web services with identical or similar functions on internet, that the existing web service selection technology based on functional description is not able to solve this problem. Therefore, the web service selection technology based on Quality of Service (QoS) with Price, response time, reputation and so on has attracted more and more people's attention recently.In this paper, the well-known classification models viz., back propagation neural network (BPNN), Bayesian network, support vector machine(SVM), classification and regression trees (CART), k-nearest neighbor and C4.5 decision tree (J48) are employed to predict the quality of a web service based on a set of quality attributes. And also a fuzzy decision tree algorithm Gini Index based (FSPRINT) is proposed to fuzzify the decision boundary without converting the numeric attributes into fuzzy linguistic terms. In order to validate the algorithm, experiments are carried out based on the QWS dataset and the 10-fold cross-validation method on the Weka platform. Experiment result shows that the FSPRINT technique outperforms all other techniques. It is found that WSRF, reliability, throughput, successability, documentation and availability are the most important attributes in that order.