Date of Award
Campus Access Dissertation
Doctor of Philosophy (PhD)
Dan A. Simovici
The demand for new techniques brought by data mining has been skyrocketting in recent years due to the massive increase in data available in many areas such as science, engineering, finance, biological research, medicine etc. We introduce several combinatorial algorithms that improve mining of binary data sets. Our contribution consists of a genetic algorithm to mine frequent itemsets and large bite sets, an algorithm that identifies determining sets for index functions, a compression algorithm that utilizes Gini-based distances between attributes in datasets, and an algorithm to find Vapnik-Chervonenkis dimension of binary tables. Also, we propose a new approach to genetic algorithms and introduce a novel type-based genetic algorithm, which we apply to two well-known problems: N-queen problem and finding the global minimum to the Rosenbrock function.
Sizov, Roman A., "Combinatorial Algorithms in Data Mining" (2019). Graduate Doctoral Dissertations. 467.