DENSITY BASED SPATIAL CLUSTERING DEVIATIONS FOR BLACK BOX REGRESSION TESTING IN LARGE DATABASE

Abstract

Ensuring the functional excellence of database applications is a most needed and important problem in software testing. Database applications are mostly adopted in many fields. For example in public administrations they need to process large amounts of transactions efficiently and need to store large amounts of data. Regression testing is used to check whether the new changes cause any errors in the existing software. Regression testing is one of the most useful software testing types during software maintenance. Clustering is a data mining technique used to discover patterns from the database. This research work incorporates soft clustering concept, which is the process of deriving the information based on the similarity of the unsupervised database. It can be considered the most important unsupervised learning technique and it deals with finding a structure in a collection of unlabeled data. Black-box Regression Testing approach focuses on regression testing and proactively exposes behavioral deviations by checking inside the black box instead of checking only black-box outputs. The proposed research work introduces a new approach called Expectation-Maximization (EM) with Density Based Spatial Clustering Algorithm to measure the dissimilarity of data elements in large database. This research work, presents an optimal perspective on the problem of EM Clustering Deviations for Black Box Regression Testing of Database Applications. The proposed method called Density based spatial clustering deviations for black box regression testing in large database which measures the correspondence between pairs of data points. The proposed method is to establish a unified framework on unsupervised data sets. To validate the proposed approach, a large scale Norwegian Tax Accounting System case study is considered, and the results show that clustering approach can indeed serve as an accurate strategy for grouping regression test deviations. The proposed analysis suggests that this approach can significantly reduce the effort spent by testers in analyzing regression test deviations, increase their level of confidence, and therefore make regression testing more scalable.