An integrated and cross-disciplinary approach towards biclustering and constraint programming

• Grant number: 20/00123-9
• Support Opportunities: Scholarships abroad - Research Internship - Post-doctor
• Start date: September 01, 2021
• End date: August 31, 2022
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Fernando José von Zuben
• Grantee: Rosana Veroneze
• Supervisor: Siegfried Nijssen
• Host Institution: Faculdade de Engenharia Elétrica e de Computação (FEEC). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Institution abroad: Universitè Catolique de Louvain (UCL), Belgium
• Associated to the scholarship: 17/21174-8 - Enumerative algorithms for biclustering: expanding and exploring their potential in bioinformatics and neuroscience, BP.PD

Abstract

Biclustering is a powerful data analysis technique that allows simultaneous clustering of both rows and columns of a data matrix. Since her PhD, Dr. Rosana Veroneze works on the development of enumerative biclustering algorithms with unique properties and functionalities that can be exploited in areas of high potential of contribution. To continue the advances in the biclustering area and produce high impact research at the forefront of science, the establishment of national and international partnerships is of great importance. For this reason, Dr. Veroneze spent three months in a technical visit to the AIA research group at UCLouvain in 2019. They are pioneers in showing that it is possible to employ constraint programming (CP) techniques for modeling and solving a wide variety of data mining tasks, including constraint-based pattern mining tasks. Remarkably, they developed a basic CP model for specifying frequent itemsets and showed that this model can easily be extended to realize the other settings of this problem. This contrasts with typical procedural data mining systems where the underlying procedures need to be modified in order to accommodate new types of constraint, or novel combinations thereof, which is usually non-trivial to be done. Moreover, the need to allow user-specified combinations of constraints is recognized in the data mining community. In its turn, biclustering problems can also be seen as constraint optimization problems. Thus, the intention of this collaboration is (1) to study the deeper integration of search and modeling strategies used in CP and biclustering; and (2) to scrutinize in detail which constraints or optimization criteria are interesting for biclustering tasks, in the context of new and existing applications. Besides to explore constraints to select interesting patterns in supervised, semi-supervised and unsupervised scenarios, we will also explore constraints to perform an automatic summarization of the results in the form of a compact description that is relevant, non-redundant and easily interpretable. We highlight that the results obtained during the technical visit are very promising and provide concrete starting points for this project. (AU)

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

DAM, KHANH HUU THE; -WILSON, THOMAS GIVEN; LEGAY, AXEL; VERONEZE, ROSANA. Packer classification based on association rule mining. APPLIED SOFT COMPUTING, v. 127, p. 21-pg., 2022-08-04. (20/00123-9, 17/21174-8)