Multiple Uses of Frequent Episodes in Temporal Process Modeling

(ABSTRACT) This dissertation investigates algorithmic techniques for temporal process discovery in many domains. Many different formalisms have been proposed for modeling temporal processes such as motifs, dynamic Bayesian networks and partial orders, but the direct inference of such models from data has been computationally intensive or even intractable. In this work, we propose the mining of frequent episodes as a bridge to inferring more formal models of temporal processes. This enables us to combine the advantages of frequent episode mining, which conducts level wise search over constrained spaces, with the formal basis of process representations, such as probabilistic graphical models and partial orders. We also investigate the mining of frequent episodes in infinite data streams which further expands their applicability into many modern data mining contexts. To demonstrate the usefulness of our methods, we apply them in different problem contexts such as: sensor networks in data centers, multi-neuronal spike train analysis in neuroscience, and electronic medical records in medical informatics. Acknowledgments It is my great pleasure to thank the people who made this thesis possible. Dr. Naren Ramakrishnan, my thesis advisor, has been my greatest source of inspiration in this entire journey which now culminates in this dissertation. Every step of the way he has provided valuable insights into the data mining problems. He has helped me understand the different subtleties of data mining methods and probabilistic models. I am deeply grateful to Dr. Srivatsan Laxman who has spent innumerable sleepless nights sharing with me valuable intuitions about the different data mining problems addressed in this dissertation. He has been a great friend and has helped with words of encouragement and advice every time I hit a wall trying to solve a problem. His positive attitude and his faith in me have been instrumental in the completion of this thesis. I want to thank Dr. Manish Marwah for giving me the opportunity to work on some of the most challenging problems in knowledge discovery. With his help and guidance we were able formulate a data mining approach to an important problem in sustainability. His eye for detail and continuous persuasion to better our methods has lead to significant improvement in the quality of this work. I am indebted to Dr. T. M .Murali and Dr. Yang Cao for their unfailing support as my thesis committee members. I am thankful for their technical insights which have helped me overcome some of …