Investigation on prototype learning.

• Other Authors: Keung, Chi-Kin.
• Format: Others
• Language: English; Chinese
• Published: 2000
• Subjects: Machine learning; Computer algorithms
• Online Access: http://library.cuhk.edu.hk/record=b5890254; http://repository.lib.cuhk.edu.hk/en/item/cuhk-323239

Keung Chi-Kin. === Thesis (M.Phil.)--Chinese University of Hong Kong, 2000. === Includes bibliographical references (leaves 128-135). === Abstracts in English and Chinese. === Chapter 1 --- Introduction --- p.1 === Chapter 1.1 --- Classification --- p.2 === Chapter 1.2 --- Instance-Based Learning --- p.4 === Chapter 1.2.1 --- Three Basic Components --- p.5 === Chapter 1.2.2 --- Advantages --- p.6 === Chapter 1.2.3 --- Disadvantages --- p.7 === Chapter 1.3 --- Thesis Contributions --- p.7 === Chapter 1.4 --- Thesis Organization --- p.8 === Chapter 2 --- Background --- p.10 === Chapter 2.1 --- Improving Instance-Based Learning --- p.10 === Chapter 2.1.1 --- Scaling-up Nearest Neighbor Searching --- p.11 === Chapter 2.1.2 --- Data Reduction --- p.12 === Chapter 2.2 --- Prototype Learning --- p.12 === Chapter 2.2.1 --- Objectives --- p.13 === Chapter 2.2.2 --- Two Types of Prototype Learning --- p.15 === Chapter 2.3 --- Instance-Filtering Methods --- p.15 === Chapter 2.3.1 --- Retaining Border Instances --- p.16 === Chapter 2.3.2 --- Removing Border Instances --- p.21 === Chapter 2.3.3 --- Retaining Center Instances --- p.22 === Chapter 2.3.4 --- Advantages --- p.23 === Chapter 2.3.5 --- Disadvantages --- p.24 === Chapter 2.4 --- Instance-Abstraction Methods --- p.25 === Chapter 2.4.1 --- Advantages --- p.30 === Chapter 2.4.2 --- Disadvantages --- p.30 === Chapter 2.5 --- Other Methods --- p.32 === Chapter 2.6 --- Summary --- p.34 === Chapter 3 --- Integration of Filtering and Abstraction --- p.36 === Chapter 3.1 --- Incremental Integration --- p.37 === Chapter 3.1.1 --- Motivation --- p.37 === Chapter 3.1.2 --- The Integration Method --- p.40 === Chapter 3.1.3 --- Issues --- p.41 === Chapter 3.2 --- Concept Integration --- p.42 === Chapter 3.2.1 --- Motivation --- p.43 === Chapter 3.2.2 --- The Integration Method --- p.44 === Chapter 3.2.3 --- Issues --- p.45 === Chapter 3.3 --- Difference between Integration Methods and Composite Clas- sifiers --- p.48 === Chapter 4 --- The PGF Framework --- p.49 === Chapter 4.1 --- The PGF1 Algorithm --- p.50 === Chapter 4.1.1 --- Instance-Filtering Component --- p.51 === Chapter 4.1.2 --- Instance-Abstraction Component --- p.52 === Chapter 4.2 --- The PGF2 Algorithm --- p.56 === Chapter 4.3 --- Empirical Analysis --- p.57 === Chapter 4.3.1 --- Experimental Setup --- p.57 === Chapter 4.3.2 --- Results of PGF Algorithms --- p.59 === Chapter 4.3.3 --- Analysis of PGF1 --- p.61 === Chapter 4.3.4 --- Analysis of PGF2 --- p.63 === Chapter 4.3.5 --- Overall Behavior of PGF --- p.66 === Chapter 4.3.6 --- Comparisons with Other Approaches --- p.69 === Chapter 4.4 --- Time Complexity --- p.72 === Chapter 4.4.1 --- Filtering Components --- p.72 === Chapter 4.4.2 --- Abstraction Component --- p.74 === Chapter 4.4.3 --- PGF Algorithms --- p.74 === Chapter 4.5 --- Summary --- p.75 === Chapter 5 --- Integrated Concept Prototype Learner --- p.77 === Chapter 5.1 --- Motivation --- p.78 === Chapter 5.2 --- Abstraction Component --- p.80 === Chapter 5.2.1 --- Issues for Abstraction --- p.80 === Chapter 5.2.2 --- Investigation on Typicality --- p.82 === Chapter 5.2.3 --- Typicality in Abstraction --- p.85 === Chapter 5.2.4 --- The TPA algorithm --- p.86 === Chapter 5.2.5 --- Analysis of TPA --- p.90 === Chapter 5.3 --- Filtering Component --- p.93 === Chapter 5.3.1 --- Investigation on Associate --- p.96 === Chapter 5.3.2 --- The RT2 Algorithm --- p.100 === Chapter 5.3.3 --- Analysis of RT2 --- p.101 === Chapter 5.4 --- Concept Integration --- p.103 === Chapter 5.4.1 --- The ICPL Algorithm --- p.104 === Chapter 5.4.2 --- Analysis of ICPL --- p.106 === Chapter 5.5 --- Empirical Analysis --- p.106 === Chapter 5.5.1 --- Experimental Setup --- p.106 === Chapter 5.5.2 --- Results of ICPL Algorithm --- p.109 === Chapter 5.5.3 --- Comparisons with Pure Abstraction and Pure Filtering --- p.110 === Chapter 5.5.4 --- Comparisons with Other Approaches --- p.114 === Chapter 5.6 --- Time Complexity --- p.119 === Chapter 5.7 --- Summary --- p.120 === Chapter 6 --- Conclusions and Future Work --- p.122 === Chapter 6.1 --- Conclusions --- p.122 === Chapter 6.2 --- Future Work --- p.126 === Bibliography --- p.128 === Chapter A --- Detailed Information for Tested Data Sets --- p.136 === Chapter B --- Detailed Experimental Results for PGF --- p.138