\addvspace {10\p@ }
\addvspace {10\p@ }
\contentsline {figure}{\numberline {2.1}{\ignorespaces Chang's algorithm for finding prototypes}}{7}
\addvspace {10\p@ }
\contentsline {figure}{\numberline {3.1}{\ignorespaces A log of some golf-playing behavior}}{9}
\contentsline {figure}{\numberline {3.2}{\ignorespaces Pseudo code for Support Based Bayesian Ranking algorithm}}{10}
\addvspace {10\p@ }
\contentsline {figure}{\numberline {4.1}{\ignorespaces Data Set Characteristics}}{12}
\contentsline {figure}{\numberline {4.2}{\ignorespaces Data Set Characteristics}}{13}
\contentsline {figure}{\numberline {4.3}{\ignorespaces Choosing the best number of features for each data set. The best choice will have a high pd along with a low pf}}{14}
\contentsline {figure}{\numberline {4.4}{\ignorespaces Example of using the cosine law to find the position of $Oi$ in the dimension $k$}}{14}
\contentsline {figure}{\numberline {4.5}{\ignorespaces Projects of points $O_i$ and $O_j$ onto the hyper-plane perpendicular to the line $O_a$$O_b$}}{15}
\contentsline {figure}{\numberline {4.6}{\ignorespaces Pseudo code for Experiment}}{16}
\contentsline {figure}{\numberline {4.7}{\ignorespaces Probability of Detection (PD) and Probability of False Alarm (PF)results}}{17}
\contentsline {figure}{\numberline {4.8}{\ignorespaces Summary of Mann Whitney U test results (95\% confidence): moving from Befroe to After.}}{18}
\contentsline {figure}{\numberline {4.9}{\ignorespaces Position of values in the 'before' and 'after' population with data set at 3, 5, 10 and 20 clusters. The first row shows the results for r=1 while the second row shows the results for r=2}}{19}
\addvspace {10\p@ }
\contentsline {figure}{\numberline {5.1}{\ignorespaces Visualization of four(4) glass forensic models}}{28}
\contentsline {figure}{\numberline {5.2}{\ignorespaces Proposed procedure for the forensic evaluation of data}}{31}
\contentsline {figure}{\numberline {5.3}{\ignorespaces PCA for iris data set}}{32}
\contentsline {figure}{\numberline {5.4}{\ignorespaces Probability of detection (pd) and Probability of False alarms (pf) using fixed values for dimensions and fixed k values for k-nearest neighbor}}{35}
\contentsline {figure}{\numberline {5.5}{\ignorespaces Pseudo code for K-means}}{36}
\contentsline {figure}{\numberline {5.6}{\ignorespaces Pseudo code for Experiment 1}}{39}
\contentsline {figure}{\numberline {5.7}{\ignorespaces Results for Experiment 1 for the 4 data sets distinguished by the number of clusters. Here for the upper and lower tables n=4 is used while r=1 is used for the upper table and r=2 for the lower table.}}{40}
\contentsline {figure}{\numberline {5.8}{\ignorespaces Pseudo code for Experiment 2}}{41}
\contentsline {figure}{\numberline {5.9}{\ignorespaces Results for Experiment 2 for the 4 data sets distinguished by the number of clusters. Here for the upper and lower tables n=4 is used while r=1 is used for the upper table and r=2 for the lower table.}}{43}
\contentsline {figure}{\numberline {5.10}{\ignorespaces Position of values in the 'before' and 'after' population with data set at 3, 5, 10 and 20 clusters. The first row shows the results for r=1 while the second row shows the results for r=2}}{44}
\contentsline {figure}{\numberline {5.11}{\ignorespaces Results for Experiment 2 of before and after results. -1 indicates that the after is better than before}}{44}
\addvspace {10\p@ }