SOFTWARE PERFOR MANCE PREDICTION USING RANDOM FOREST BASED REGRESSION AN AL YSISJournal: International Journal of Engineering Sciences & Research Technology (IJESRT) (Vol.5, No. 10)
Publication Date: 2016-10-30
Authors : R. Sathya; P. Sudhakar;
Page : 288-293
Keywords : performance pred iction; random forest; categorical data; feature interaction; regression analysis;
The evaluation of various software quality metrics like performance, reliability, and response time are done using quantitative techniques and it is essential for component based software applications. In this paper the performance of the software applicat ion is predicted using regression analysis. In general, the trend analysis technique is employed to predict the performance of the software system. The proposed method in this paper will help the users of the software system to predict whether it satisfies their requirements for a set of features selected by them. The performance of the software gets vary based on the features selected by the users. The features may interact with other feature and degrade the overall performance of the system. The performan ce prediction is carried out using the Random forest which is capable of handling thousands of input variables without deleting the variables. Also it offers an experimental method for the detection of the feature interactions.
Other Latest Articles
Last modified: 2016-10-15 20:49:01