Web service classification and prediction using rule-based approach with recommendations for quality improvements

Abstract

Web-based applications are more popular and are increasing the demand to use of applications. Service quality and user satisfaction are the most significant for the designer. IT provides the best solutions for various applications such as B2B, B2C, e-commerce and other applications. The client demands high-quality service, regarding minimum response time, high availability and more security. The existing approaches and models may not improve the overall performance of web-based application due to not covering all functional, nonfunctional parameters. The proposed model-approach gives the best solution using rule-based classification of web services using QWS dataset, and predictions of quality parameters based on user specifications. The model is implemented in Java, the results of quality parameters will classify and predict the class labels Class-1(high quality), Class-2, Class-3 and Class-4 (low quality), and the system will give recommendations to improve the quality parameters. By using suggested guidelines and instructions to the software developer, that he will meet the client specifications and provides best quality values which will improve the overall performance (in specification parameters including functional and nonfunctional values). The result clearly suggests the improvement of quality parameters by classification and prediction. This paper can be extended to mixed attributes of quality parameters.