DNA copy number variations are now recognized as an important contributor to human genetic disease. In this paper, our focus is on identifying patterns of DNA copy number variation detected with finely-tiled oligonucleotide arrays in casecontrol studies. This analysis is based on the observation that CNVs across large segments of the genome show recurring patterns, particularly in regions that bear repeat sequences that contribute to the genetic instability of that interval. The goal of this analysis is to increase the power to identify diseaseassociated genetic changes in case-controls studies of copy number variation. We propose a framework to evaluate the predictive power of recurrent variations at multiple genomic sites. First, we present a novel method based on maximum likelihood principle to clearly map and detect copy number variations along the studied genomic segments. Second, we apply regional analysis to evaluate the statistical and biological significance of recurrent variations followed by clustering methods to classify the tested samples. Finally, our results show that using the concatenated recurrent variant regions will considerably increase classification performance when compared with the traditional classifiers that use the entire data set. The results also provide insight into the pattern of the variations that may have a direct role in the targeted disease and can be used to improve diagnostic reliability for complex human genetic disorders.