Title Page
Abstract
국문 초록
Contents
Chapter 1. Introduction 11
1.1. Network Analysis 13
1.2. Gaussian Process 14
1.3. Paper outline 15
Chapter 2. ABIDE dataset 16
2.1. Demographic Information 16
2.2. Preprocessing 18
2.3. FC network construction 19
2.4. Exploratory Data Analysis 20
Chapter 3. Models 25
3.1. GP-probit and GP-logistic model 26
3.2. I-probit and I-logistic model 28
3.2.1. I-prior model 29
3.2.2. I-probit and I-logistic model with network inputs 30
Chapter 4. Implementation 34
4.1. GP-probit and GP-logistic Model 34
4.2. I-probit and I-logistic Model 36
Chapter 5. Simulation Studies 37
5.1. Simulation Study 1 37
5.2. Simulation Study 2 40
Chapter 6. Real Data Analysis 45
Chapter 7. Conclusion 48
Reference 50
Appendices 56
Appendix A. Simulation 1 results (logistic) 56
Appendix B. Simulation 2 results (logistic) 58
Table 1. Demographic information for 120 subjects 17
Table 2. Network summary statistics for ASD and control 23
Table 3. Real data classification result for m=120 subjects (GP-probit, GP-logistic models). 46
Table 4. Real data classification result for m=120 subjects (I-probit, I-logistic models). 46
Figure 1. Examples of individual FC networks 21
Figure 2. Examples of top 20 ROI connections 22
Figure 3. Simulation 1 results over 100 replicates for each 3 distance options. 38
Figure 4. Simulation 2 results of GP-probit model over 10 replicates for each 5 random graph models. 42
Figure 5. Simulation 2 results of I-probit model over 10 replicates for each 5 random graph models. 43