\contentsline {chapter}{\numberline {1}Introduction}{1}
\contentsline {section}{\numberline {1.1}Motivation}{2}
\contentsline {section}{\numberline {1.2}Statement of Thesis}{3}
\contentsline {section}{\numberline {1.3}Contributions}{3}
\contentsline {section}{\numberline {1.4}About This Document}{4}
\contentsline {chapter}{\numberline {2}Background: Data and Learners}{5}
\contentsline {section}{\numberline {2.1}Data and Data Mining}{5}
\contentsline {subsection}{\numberline {2.1.1}Data}{5}
\contentsline {subsection}{\numberline {2.1.2}Data Mining}{7}
\contentsline {section}{\numberline {2.2}Classification}{8}
\contentsline {subsection}{\numberline {2.2.1}Decision Tree Learners}{9}
\contentsline {subsection}{\numberline {2.2.2}Naive Bayes}{13}
\contentsline {subsection}{\numberline {2.2.3}Other Classification Methods}{16}
\contentsline {subsubsection}{1-R}{16}
\contentsline {subsubsection}{Rule Learners}{18}
\contentsline {subsubsection}{Instance-Based Learning}{20}
\contentsline {section}{\numberline {2.3}Summary}{21}
\contentsline {subsection}{\numberline {2.3.1}Data Mining and Classification}{21}
\contentsline {subsection}{\numberline {2.3.2}Classifier Selection}{22}
\contentsline {chapter}{\numberline {3}Discretization}{24}
\contentsline {section}{\numberline {3.1}General Discretization}{24}
\contentsline {section}{\numberline {3.2}Equal Width Discretization (EWD)}{28}
\contentsline {section}{\numberline {3.3}Equal Frequency Discretization(EFD)}{29}
\contentsline {section}{\numberline {3.4}Bin Logging}{30}
\contentsline {section}{\numberline {3.5}Entropy-based Discretization}{30}
\contentsline {section}{\numberline {3.6}Proportional k-Interval Discretization}{32}
\contentsline {section}{\numberline {3.7}Weighted Proportional k-Interval Discretization (WPKID)}{35}
\contentsline {section}{\numberline {3.8}Non-Disjoint Discretization (NDD)}{35}
\contentsline {section}{\numberline {3.9}Weighted Non-Disjoint Discretization (WNDD)}{36}
\contentsline {section}{\numberline {3.10}Other Methods}{37}
\contentsline {section}{\numberline {3.11}DiscTree Algorithm}{37}
\contentsline {subsection}{\numberline {3.11.1}Trees}{38}
\contentsline {subsection}{\numberline {3.11.2}Binary Trees}{40}
\contentsline {subsection}{\numberline {3.11.3}Binary Search Trees}{42}
\contentsline {subsection}{\numberline {3.11.4}Randomized Binary Search Trees}{45}
\contentsline {subsection}{\numberline {3.11.5}DiscTree}{46}
\contentsline {chapter}{\numberline {4}Experiment}{51}
\contentsline {section}{\numberline {4.1}Test Data}{51}
\contentsline {section}{\numberline {4.2}Cross-Validation}{53}
\contentsline {section}{\numberline {4.3}Classifier Performance Measurement}{54}
\contentsline {section}{\numberline {4.4}Mann-Whitney}{56}
\contentsline {chapter}{\numberline {5}Experimental Results}{60}
\contentsline {section}{\numberline {5.1}DiscTree Variant Selection}{60}
\contentsline {subsection}{\numberline {5.1.1}Accuracy Results}{61}
\contentsline {subsection}{\numberline {5.1.2}Balance Results}{67}
\contentsline {subsection}{\numberline {5.1.3}Precision Results}{73}
\contentsline {subsection}{\numberline {5.1.4}Probability of Detection Results}{79}
\contentsline {subsection}{\numberline {5.1.5}Probability of Not False Alarm}{85}
\contentsline {subsection}{\numberline {5.1.6}Decision Tree Method Selection}{91}
\contentsline {section}{\numberline {5.2}Discretization Method Comparison}{91}
\contentsline {subsection}{\numberline {5.2.1}Accuracy Results}{91}
\contentsline {subsection}{\numberline {5.2.2}Balance Results}{98}
\contentsline {subsection}{\numberline {5.2.3}Precision Results}{105}
\contentsline {subsection}{\numberline {5.2.4}Probability of Detection Results}{112}
\contentsline {subsection}{\numberline {5.2.5}Probability of Not False Alarm}{119}
\contentsline {section}{\numberline {5.3}Summary}{126}
\contentsline {chapter}{\numberline {6}Conclusion}{127}
\contentsline {section}{\numberline {6.1}Overview}{127}
\contentsline {section}{\numberline {6.2}Conclusions}{128}
\contentsline {section}{\numberline {6.3}Future Work}{129}
\contentsline {chapter}{\numberline {A}disctree Source Code}{132}
\contentsline {chapter}{\numberline {B}crossval Source Code}{141}
\contentsline {chapter}{\numberline {C}tenbins Source Code}{143}
\contentsline {chapter}{\numberline {D}Script for PKID}{144}
\contentsline {chapter}{\numberline {E}Entropy-Minimization Method Script}{145}
\contentsline {chapter}{\numberline {F}Performance Measure $U$-test Tables}{146}
\contentsline {section}{\numberline {F.1}Accuracy $U$-test By Data Set}{146}
\contentsline {section}{\numberline {F.2}Balance $U$-test by Data Set}{151}
\contentsline {section}{\numberline {F.3}Precision $U$-test by Data Set}{156}
\contentsline {section}{\numberline {F.4}Probability of Detection $U$-test by Data Set}{161}
\contentsline {section}{\numberline {F.5}Probability of Not False Alarm $U$-test by Data Set}{166}