Vertical Association Rule Mining

This work focuses on the data-mining task of §association rule mining which discovers association §relationships among items in datasets matching user-§defined measures of interest. We describe an §efficient vertical framework for representing data §and mining frequent itemsets that is based on the P-§tree technology along with other artificial §intelligence techniques, such as set-enumeration §trees and tabu search. With the objective of §handling the mounting needs of many applications, §such as precision agriculture, the proposed §framework is used to produce rules in situations §where the ubiquitous support-based pruning is not §sought. In the context of citation graphs, our §proposed framework operates in a (semi) divide-and-§conquer parallelized fashion, to discover patterns §among subject matters that reveal the evolution §history and any possible future extensions of §subject matters. The same framework is utilized §in an interactive incremental parallel model which §focuses on analyzing genome annotation data for §association rules potentially useful in annotating §new genes, replacing missing values, and validating §old annotations. This book provides a unique and fairly comprehensive treatment of a popular data mining task known as association rule mining. At the heart of this book, a vertical framework (based on the patented P-tree technology along with other well-known artificial intelligence techniques) for data representation and data mining is described. The framework is adapted to environments that require divide and conquer parallel processing or pruning beyond the ubiquitous support-based pruning. Along with the theory, the book highlights the versatility of the presented framework in different application domains ranging from citation analysis to precision agriculture and bioinformatics.