Efficient Techniques for Decision Support: Approximate Query Processing to Parallel and Value-Based Data Mining.

Abstract

Modern techniques of capturing data have thrown, besides storage, another couple of challenges to the computer scientists, viz. its quick retrieval and efficient processing. Getting the information quickly in today’s ever-increasing data deluge is a key priority for the decision maker. This text examines and describes some new structures and techniques in this area. The purpose of this research is to investigate efficient techniques including data structures, algorithms and their implementations for decision support applications in data warehousing and data mining. The specific techniques proposed include a new efficient indexing structure for approximate query processing, a parallel algorithm for mining frequent patterns, and the mining of value-based itemsets by finding optimal solutions under resource constraints. The effectiveness of each technique has been evaluated using typical test data sets. Written both for computing and information systems researchers, this text is aimed at advanced researchers, particularly, in the area of data warehousing and data mining and, in general, for the database professionals who are keen to know about efficient data organisation.