\contentsline {chapter}{\numberline {1}Introduction}{1}
\contentsline {section}{\numberline {1.1}Statement of Thesis}{4}
\contentsline {section}{\numberline {1.2}Contributions of this Thesis}{4}
\contentsline {section}{\numberline {1.3}Structure of this Thesis}{4}
\contentsline {chapter}{\numberline {2}Background and Related Work}{5}
\contentsline {section}{\numberline {2.1}Prototype Learning for Nearest Neighbor Classifiers}{5}
\contentsline {subsection}{\numberline {2.1.1}Instance Selection}{6}
\contentsline {subsubsection}{Condensed Nearest Neighbor (CNN)}{6}
\contentsline {subsubsection}{Reduced Nearest Neighbor (RNN)}{6}
\contentsline {subsubsection}{Minimal Consistent Set (MCS)}{6}
\contentsline {subsection}{\numberline {2.1.2}Instance Abstraction}{6}
\contentsline {subsubsection}{Chang}{6}
\contentsline {chapter}{\numberline {3}CLIFF: Tool for Instance Selection}{8}
\contentsline {section}{\numberline {3.1}CLIFF: Tool for Instance Selection}{8}
\contentsline {chapter}{\numberline {4}CLIFF Assessment}{11}
\contentsline {section}{\numberline {4.1}Data and Preprocessing Tools}{11}
\contentsline {subsection}{\numberline {4.1.1}Data Set Characteristics}{11}
\contentsline {subsection}{\numberline {4.1.2}Pre-processing tools for Dimensionality Reduction}{12}
\contentsline {subsubsection}{FastMap}{12}
\contentsline {subsubsection}{Feature Subset Selection (FSS)}{13}
\contentsline {section}{\numberline {4.2}CLIFF Assessment on Standard Data Sets}{15}
\contentsline {subsection}{\numberline {4.2.1}Experimental Method}{15}
\contentsline {subsection}{\numberline {4.2.2}Is CLIFF viable as a Prototype Learner for NNC?}{16}
\contentsline {subsubsection}{Results from Experiment}{17}
\contentsline {chapter}{\numberline {5}Case Study: Solving the Problem of Brittleness in Forensic Models}{20}
\contentsline {section}{\numberline {5.1}Introduction}{20}
\contentsline {section}{\numberline {5.2}Visualization of Brittleness}{22}
\contentsline {section}{\numberline {5.3}Glass Forensic Models}{22}
\contentsline {subsection}{\numberline {5.3.1}Seheult 1978}{23}
\contentsline {subsection}{\numberline {5.3.2}Grove 1980}{24}
\contentsline {subsection}{\numberline {5.3.3}Evett 1995}{25}
\contentsline {subsection}{\numberline {5.3.4}Walsh 1996}{26}
\contentsline {section}{\numberline {5.4}Visualization of Brittleness in Models}{27}
\contentsline {section}{\numberline {5.5}Introduction}{29}
\contentsline {section}{\numberline {5.6}Dimensionality Reduction}{31}
\contentsline {subsection}{\numberline {5.6.1}Principal Component Analysis}{31}
\contentsline {section}{\numberline {5.7}Clustering}{34}
\contentsline {section}{\numberline {5.8}Classification with KNN}{34}
\contentsline {section}{\numberline {5.9}The Brittleness Measure}{36}
\contentsline {section}{\numberline {5.10}Data Set and Experimental Method}{37}
\contentsline {section}{\numberline {5.11}Experiment 1: KNN as a forensic model?}{38}
\contentsline {subsection}{\numberline {5.11.1}Results from Experiment 1}{38}
\contentsline {section}{\numberline {5.12}Experiment 2: Can brittleness be reduced?}{39}
\contentsline {subsection}{\numberline {5.12.1}Results from Experiment 2}{41}
\contentsline {chapter}{\numberline {6}Conclusion}{45}