In recent years, an increasing number of important experimental problems require the collection and the analysis of data about a large number of features about the system under study. For example, protein engineering constructs new proteins by examining combination of several amino-acids into sequences. In genetics, researchers study the simultaneous behaviour of several genes from micro-array data. Many social systems, including those in economics and finance, are described by complex networks of interactions among several actors. Our group in Evolutionary Design of Experiments and Data Mining develops a variety statistical methods to collect, model and analyze high-dimensional data. One important focus of our research is the development of data-gathering techniques to improve experimental results and reduce the time or cost of experimentation. In the past, we developed several adaptive methods based on the paradigm of the natural evolution such as evolutionary neural network, ant colony optimization and evolutionary Bayesian networks. Currently, we are developingnovel methods for adaptive collection of data when dealing with complex experiments in high dimensions. In our approaches, we merge the strengths of the state of the art from different areas: evolutionary computing, stochastic optimization, information theory, statistics and probability theory.

Another important area of research is the development statistical model-building techniques for making accurate predictions in high-dimensions. Our research include methods for selection/combining of statistical models, function estimation, robust statistical procedures and methods for classification.