Browse Theses

This thesis presents some exploration in the field of data mining. Data mining is popularly referred to as knowledge discovery in databases (KDD), and is the automated or convenient extraction of patterns representing knowledge implicitly stored in databases, data warehouses, and other massive information repositories. This thesis explores association rule and quantitative association rule mining among infrequent items in the field of data mining.

Association rule mining, playing a critical role in the field of data mining, searches for interesting relationships among items in a given data set. Association rule mining among frequent items has been extensively studied in data mining research. However, in the recent years, there is an increasing demand of mining the infrequent items (such as rare but expensive items). Since exploring interesting relationship among infrequent items has not been discussed much in the literature, in this thesis, we propose two practical and effective schemes, Matrix-Based Scheme and Hash-Based Scheme, to mine association rules among rare items. These two methods can also be applied to efficiently capture interesting association patterns among frequent items with bounded length. Experiments are conducted to test behaviors of our algorithms.

Quantitative association rule mining has been mainly studied in relational database. In this thesis, we explore quantitative association rule mining in relational database among infrequent items. We reanalyze association rules with quantity incorporated. Experiments are drawn to illustrate the more interesting and informative rules captured.