Research Statement

My thesis draft is available now. All files have two copies in gzipped postscript format and (un-gzipped) pdf format.

Thesis

The complete dissertation in one file:

    thesis.ps.gz   (178 pages) and [PDF]

Or, each chapter its own file:


  OMEGA: On-line MEmory-based GenerAl purpose system detector  

 FRONTPAGES   [PDF]

 CONTENTS   [PDF]

 CHAPTER 1  Introduction  [PDF]
            1.1  What is system classification?
            1.2  The applications of system classification
            1.3  The assumptions of OMEGA
            1.4  Related fields
            1.5  The system classification approaches
            1.6  Thesis outline
            1.7* A brief introduction to HMM

 CHAPTER 2  Memory-based System Detection [PDF]
            2.1  Likelihood analysis
            2.2  Hypothesis testing
            2.3  Efficiency issues
            2.4  Pre-processing
            2.5  Memory-based learning
            2.6  Summary

 CHAPTER 3  Tennis Style Detection [PDF]
            3.1  Experimental design
            3.2  OMEGA result analysis
            3.3  Comparison with other methods
            3.4  Summary

 CHAPTER 4  Logistic Regression as a Classifier [PDF]
            4.1  Classification methods
            4.2  Global logistic regression
            4.3  Locally weighted logistic regression
            4.4  Comparison experiment
            4.5  Summary

 CHAPTER 5  Efficient Memory Information Retrieval  [PDF]
            5.1  Efficient information retrieval
            5.2  Kd-tree construction and information retrieval
            5.3  Cached kd-tree for memory-based learning
            5.4  Experiments and results
            5.5. Summary

 CHAPTER 6  Kd-tree for Various Regressions [PDF]
            6.1  Locally weighted linear regression
            6.2  Efficient locally weighted linear regression
            6.3  Technical details
            6.4  Empirical evaluation
            6.5  Kd-tree for logistic regression
            6.6  Empirical evaluation
            6.7  Summary

 CHAPTER 7  Feature Selection [PDF]
            7.1  Introduction
            7.2  Cross-validation vs. Overfitting
            7.3  Feature selection algorithms
            7.4  Experiments
            7.5  Summary

 CHAPTER 8  Driving Simulation  [PDF]
            8.1  Experiment data
            8.2  Experiment results
            8.3  Comparison with other methods
            8.4  Summary

 CHAPTER 9  Real World Driving  [PDF]
            9.1  Data collection
            9.2  OMEGA result
            9.3  Comparison with other methods
            9.4  Summary 

 CHAPTER 10  Conclusion [PDF]
            10.1  Discussion
            10.2  Contributions
            10.2  Future research
            10.3  Applications

 APPENDIX A  Chinese Handwriting Recognition [PDF]

 REFERENCES  [PDF]

Kan Deng

Last modified: Tuesday April 13, 1999

Back to homepage.