CS Student Presented in Top Data Science Conference
Jarrod Moore, a Computer Science Masters Student, has been conducting research in cluster analysis for his thesis under the guidance of Dr. Ackerman and Dr. Gary Tyson. His recent work on perturbation robustness in clustering has been published in a paper titled “Foundations of Perturbation Robust Clustering” in ICDM 2016, a top tier conference in the area of Data Science and Machine Learning. The paper is co-authored by Dr. Ackerman at San Jose State University.
In the paper, Jarrod Moore and Dr. Ackerman analyze the effects of perturbations, small changes to existing data points, on all possible clustering algorithms. Formally proving that robustness (resistance to large changes) is impossible in the presence of perturbations in the general case. Further, the paper provides analysis of common clustering techniques, providing what criteria must be met in order to ensure robustness to perturbations.