Dr Marco Caserta is a Visiting Fellow from IE University Madrid, and will be at CRASSH in Lent and Easter Terms 2018.
Learning from data is one of the most challenging tasks of modern science. To deal with the growing amount of available data, researchers in different fields such as statistics, data mining, and engineering have focused on the design and development of efficient algorithms for large datasets. The application of learning algorithms spans fields as diverse as medicine, genetics, finance, and linguistic, among others.
My interest lies in the development of learning algorithms for clustering and classification of large datasets. More specifically, I am working on the development of a novel data classification algorithm based on Logical Analysis of Data (LAD), a combinatorial optimization-based approach. I
am currently applying such novel classification method to sentiment analysis, i.e., the classification of tweets, and the performance of this newly designed algorithm are competitive with traditional approaches for data classification.
Potential developments of my line of research in connection with the project carried out at the CRASSH Concept Lab include the application of classification and clustering algorithms to the analysis of large datasets of texts, e.g., using unsupervised techniques to find clusters and patterns within the texts. My contribution to the project is mainly computational, focused on the design and implementation of clustering and
classification algorithms and other machine learning tools to identify trends, relations, and categories within large datasets.
I am a Professor of Statistics at IE University and IE Business School, Madrid, Spain. I hold a Ph.D. in Operations Research from the University of Illinois, Chicago, USA (2004) and a MS in Management Engineering from Politecnico di Milano, Italy (2000). I am a researcher in the field of Operations Research and Management Science, an area at the intersection of applied mathematics, computer science, and management.
My research up to date has covered both the theory and the application of operations research techniques. From the theoretical perspective, I focus on the properties of so called “matheuristic” algorithms, algorithms that exploit classical mathematical programming techniques in a heuristic fashion. These algorithms are typically used in real-world settings, where pure optimization techniques would fail to find solutions in a reasonable amount of time. From the application point of view, I tackled problems in collaboration with a number of companies and universities, with a keen interest in testing the applicability of the theoretical advancements we developed.
My main contributions and findings are summarized in over 40 peer reviewed publications, among others 20 papers in international journals, 4 book chapters, 1 edited book, and a number of conference papers. I was also the co-organizer of an international conference, the MIC 2009 edition of
the Metaheuristic International Conference, and co-edited the post-conference book, appeared under the Springer Series in Operations Research. My trajectory as researcher has been recognized multiple times by the Spanish Commission for the Evaluation of Research Activities (CNEAI).