Software architecture is playing a crucial role in software projects. However, many software projects do not have a software architecture or their architecture is out of date. For these reasons, it has become important to recover the software architecture nowadays. In the literature, many methods have been presented to recover software architecture.These methods are not usable because they depend on just one cluster rather than cluster ensembles. We believe that leveraging cluster communities will be of paramount importance in software architecture recovery. Cluster aggregation is a powerful technique for improving the accuracy and robustness of clustering results. By combining the output of multiple clustering algorithms, it is possible to mitigate the effects of noise and outliers and produce more reliable clusters. In this report, we propose a clustering aggregation algorithm that is based on the idea of majority voting. Given the output of multiple clustering runs, our algorithm first computes a Design Structure Matrix (DSM) that captures the degree of overlap between each pair of clusters. It then applies a clustering aggregation technique to this matrix in order to obtain a new set of clusters that reflects the consensus of the original clustering runs. We evaluate the performance of our algorithm on a variety of datasets. Our experiments show that our algorithm adequately performs compared to the existing approaches in terms of both clustering accuracy and computational efficiency. Overall, our results suggest that the proposed algorithm is a promising tool for improving the reliability of clustering results in a variety of applications.