Date of Award
Campus Access Dissertation
Doctor of Philosophy (PhD)
Dan A. Simovici
Data mining is used in many areas of science and engineering, such as bioinformatics, genetics, finance and electrical engineering. The increased used of data mining brings a lot of new techniques, but with the cost of complexity and specialized problems. In electrical engineering, automated design of circuits is challenging due to the growing scale and complexities of the circuits. Data mining techniques improve the existing solutions, allowing circuit design to be fully automated or with minimal human intervention. Our contribution consists of a method that detects the minimal sets of variables that determine the values of a discrete partially defined function, and in a novel method of decomposition of partially specified index generation functions (PSIGFs). The data mining process can be costly. Therefore, it is important to evaluate the potential payoff of the mining process before the actual mining takes place. We propose a new approach for evaluating the minability of data sets by using compression. The basic idea is that compressible data contains patterns and the existence of these patterns makes the data worth mining.
Pletea, Dan Alexandru, "Efficient Mining Algorithms in Engineering" (2013). Graduate Doctoral Dissertations. 135.