Research And Applications Of Data Mining

As an interdisciplinary field, Data Mining (DM) combines the efforts of statistics, computer science, pattern recognition, artificial intelligence, machine learning and other disciplines together. In today's digital society, the explosive growth of many business, government, and scientific databases have far outpaced our ability to interpret and digest this data, creating a need for a new generation of tools and techniques for automated and intelligent database analysis, and that is the goal of DM.We focus on following several aspects in DM: decision trees construction; the attribute selective criterion of decision trees algorithm; rough set theory and application, agent based DM system; and a general multi-tasks oriented DM system and its integrated method.The main work of the thesis includes:1. A decision trees algorithm BSDT based on bias shift is given. The bias shift is complemented by the means of a two-tier algorithm, which is based on case-based reasoning. The first tier is aimed to select a learning algorithm which is suited the goal task best. The training set of the first tier is constructed based on typical examples and existed algorithm, and the learning algorithm is GSD. The output algorithm of the first tier is the input learning algorithm of the second tier. The primitive training set is used as training set of the second tier. The target classification decision tree and/or rules are obtained. In addition to, a mechanism to get new algorithm and typical examples are given.2. The background knowledge of the given task such as preference, cost and conceptual hierarchy is used. The generalization algorithm and specialization algorithm is taken advantage of to preprocess the primitive train sets so as to get a designated conceptual hierarchy. Meanwhile, an attribute selective function ASF is obtained in terms of preference coefficient and cost coefficient.3. A notion of a similitude matrix and association matrix as the extended versions of discernibility matrix is presented, which is an important...