WS COMPONENT SELECTION BY IMPROVISED HIGH HIT RATIO USING SIMPLE JACCARD COSINE DISTANCES WITH MODI'S COST EFFECTIVENESS

Software component is an inevitable commodity in the field of web technology applications. Any business transaction in online has been taken care by the software component as the whole. Web service is a software component, which is all articulating highly in the market. Selection and prediction for such a type of component is a tough task for our application. Prism classifier is a statistical tool through which obtaining good classification and ontology for our semantics with number of attributes. Every web service component has its own significance and QOS. Prism classifier generates output considering only high values, resulting in the rules, which contains only the best component, skipping the next components in the priority queue. The drawback in classical prism classifier is rectified by considering the attributes of the component having tie between maximum values. The homogeneity levels amongst a class, variation between the training data sets are also analyzed. By, improving the prism classifier, the resulting rule contains the best of the best component suitable for the customer. The series of tests like Simple Matching co-efficient, Jaccard distances, cosine distance, T-test, ANOVA etc., together with modified Prism classifier is named as IH2RC [Improvised High Hit Ratio Classifier]. In this paper, IH2RC is applied on a training data set, which contains online translators with their related attributes. For cost effectiveness of the software component MODI'S method is employed in this scenario.