MSc Defence – Jai Liu

The School of Computer Science is pleased to announce the following MSc Defence, Using Association Rules and Logical Learning for Clustering, presented by MSc Candidate Jia Liu.

The defence will take place August 21, 2014 in Reynolds 219 at 10:30 am.

Title

Using Association Rules and Logical Learning for Clustering

Abstract

The well-known a priori algorithm generates a set of so-called association rules from frequent tuples or item sets. These association rules are far from a complete logical description of what, in our case, constitutes membership in two or more clusters of data. In this thesis, it is proposed to edit and reduce the association rules from an apriori analysis to a consistent and compact form using a current best hypothesis (CBH) logical learning scheme. The procedure is tested on a large, but very sparse dataset describing the relationship of agricultural and construction activity on the macroinvertebrate populations in the surface waters of the various sub-basins of the Liaoning Basin of the Northeastern part of the Peoples Republic of China. This thesis also includes the comparison of the association rules / CBH results with K- means method and results (generated by the commercially available PC-ord software). Previous work has tested the hypothesis that a threshold of 21% or more of land area in a watershed is the tipping point to surface water degradation. From their analysis, a logical system is constructed which could then be used to further understand the relationships of agriculture and construction land use and mitigation procedures such as riparian barriers to intercept and treat runoff.

Advisor: Fei Song / David Swayne