Concurrent and Spectral Clustering of Wireless Waves
Unsupervised clustering is generally used to identify groupings of wireless waves from estimated multipath parameters in order to optimize the cluster count and membership. The estimated parameters exclude cluster labels. For a better comparison of clustering approaches, cluster-based wireless channel models provide the labels. This work proposes that using the three-constraint affinity matrix (3CAM) in formulating the affinity or similarity matrix improves the clustering accuracy. Datasets generated from the European Cooperation in Science and Technology (COST) 2100 channel model (C2CM) were used and subjected to directional cosine and whitening transforms. Simultaneous clustering and model selection matrix affinity (SCAMSMA), 3CAM-SCAMSMA, spectral clustering (SC), and 3CAM-SC were used to concurrently determine cluster count and membership. Various studies on multipath clustering give only the number of clusters. Others would state only the validity index of the membership of clusters. The problem with such an approach is that the correctness of the number of clusters is not an assurance that the membership of the clusters is accurate. The four clustering approaches solve this problem by determining the number of clusters and their membership. Thus, knowing each technique’s performance is essential. In the algorithms of all the clustering approaches, cluster count aims to ensure that the target cluster count is within the vicinity of the reference clusters. The cluster count and membership accuracy are computed through the cluster-wise Jaccard index of the multipath membership to their clusters. The performance of the clustering approaches was validated using the Jaccard index by comparing the calculated data with the reference data.
Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in any component of these datasets/codes are trade names, service marks, trademarks or registered trademarks of their respective owners. The author(s)/publisher(s)/authorizing body is/are not associated with any product or vendor mentioned in any component of these datasets/codes.
The inclusion of an organization name, product, or service in any part of these datasets/codes should not be construed as an endorsement of such organization, product, or service, nor is failure to include an organization name, product, or service to be construed as disapproval.
Any component or publication of these datasets/codes is designed to provide accurate and authoritative information in regard to the subject matter covered. Every attempt has been made to ensure accuracy at the time of publication. Any component or publication of these datasets/codes is made on the understanding that the the author(s)/publisher(s)/authorizing body is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought.
Any statements expressed in any component or publication of these datasets/codes are those of the individual authors and do not necessarily represent the views of the authors' affiliations, which takes no responsibility for any statement made herein. No reference made in this publication to any specific method, product, process, or service constitutes or implies an endorsement, recommendation, or warranty thereof by the authors' affiliations. The materials are for general information only and do not represent a standard of the authors' affiliations, nor are they intended as a reference in applicable specifications, contracts, regulations, statutes, or any other legal document. The authors' affiliations makes no representation or warranty of any kind, whether express or implied, concerning the accuracy, completeness, suitability, or utility of any information, apparatus, product, or process discussed in this publication, and assumes no liability therefor. The information contained in these materials should not be used without first securing competent advice with respect to its suitability for any general or specific application. Anyone utilizing such information assumes all liability arising from such use, including but not limited to infringement of any patent or patents.