Checking Language Dependent Accuracy of Web Applications using Data Mining Techniques

Over the last decade web applications are becoming very popular. These are becoming more users oriented now days. Various languages used for the development of a web application like PHP, Java, ASP.NET etc. Development of a web application is not done by individual; it is a result of team’s efforts. Different type of bugs and errors are present in source code. Finding out these bugs or errors is a difficult task. Deep understanding of the language is required to detect bugs or errors in source code. Different tools are used to check the accuracy of the source code. There is a need to classify these detected bugs so that fewer efforts are done for selecting the individual to correct these. To detect and classify these manually is a time consuming task. This paper has presented the combination of Software Engineering with Data Mining Techniques. Aim of this work is to detect and classify the bugs or errors present in source code of web applications developed using PHP language. The input to system is PHP source code file. Bugs and errors are detected after that clustering and classification is performed for knowledge discovery. Experimental results show that given work is more effective than the previous method on the basis of accuracy, recall, precision and F-Measure. Keywords— Software Engineering, Data Mining, PHP, Web Application, Bug, Error, Modified K-Means, CART