Grant: Discovery Grants Program - Individual

My research program aims to build an intelligent system that assists human to make informed and wise decisions on complex problems involving multi-criteria and multiple agents. There remain two major challenges, too many choices and contradictive criteria, in many decision-making problems especially when large, complex, unstructured, and incomplete data are involved. The proposed research program aims to utilize machine learning methods for complex and critical decision makings. In particular, I will examine theories and applications of game-theoretic learning including game-theoretic rough sets and game-theoretic shadowed sets. In addition, I will examine theories and applications of three-way clustering and aim to build game-theoretic three-way clustering models. Applications will be drawn from text classification and image processing. It is hoped that this research will broaden our knowledge on intelligent decision support mechanisms. My long-term goal is to investigate the theoretic and practical issues of machine learning, to explore the governing mechanisms of learning methods, to examine relationships between various machine learning techniques, and to incorporate such techniques into decision making processes. In particular, I will mainly focuses on descriptive analytics and predictive analytics. My short-term objectives are to study learning mechanisms and the applicability of some machine learning methods such as classification and clustering. This  program  proposes to utilize machine learning methods, such as rough sets and shadowed sets, incorporated with game theory to form a game-theoretical learning mechanisms for complex data analyses and decision making. There are four short-term goals in this research program: to further examine the theory and applications of game-theoretic rough set models with different game settings such as cooperative game and Pareto optimization; to examine the theory and applications of game-theoretic shadowed sets; to formulate game-theoretic learning based on studies of game-theoretic rough sets and game-theoretic shadowed sets;  to examine the theory and applications of three-way clustering and to investigate the possibility to incorporate game theory to three-way clustering algorithms. This work is novel in that it seeks to provide a mechanism to solve two major decision-making problems. It is expected to be significant to computer science because the game-theoretic approach for three-way classification and three-way clustering is new and relevant for various applications. It is also hope that the work will broaden our knowledge on intelligent decision support mechanisms.