Adviser : Prof. M.L. Lin

1
Potential Analysis of Debris Flow Torrents

• Adviser Prof. M.L. Lin
• Student Yu-Hsiang Lin
• Date 2007/11/23

2
Outlines

• Introduction
• Objectives
• Methodology
• PCA Fishers discriminate analysis
• new factor
• Things To Do

3
Introduction

• Every year typhoons cause disaster in Taiwan,
• and the debris flow is one of the most serious
• ground disaster.
• If we can understand the most main influence
• factor, it will provide information for disaster
• reduction management.

4
Objectives

• Principal component analysis of application
• by the geological classification
• Search for new factor and other discriminate
  analysis of multivariate statistical analysis

5
Flow Chart
Proceeding
Not yet
6
Literature Survey
7
Literature Survey
8
Study Area Database
9
Database
10
Common Factors

• Basin Area BA
• Stream Length SL
• Hypsometric Integral HI
• Stream Mean Slope SMS
• Form Factor FF
• Slope Aspect SA (N,NE,E,SE,S,SW,W,NW)
• Slope Distribution SD (SD10,SD15,SD20,
• 
  SD30,SD45,SD90)
• GEO (E1,E2,EO,Mj,Ml,MS,O1,O2,Q2,Q3,Qo)

11
Principle Component Analysis

• Principle component analysis (PCA) is a
• technique used to reduce multidimensional
• data sets to lower dimensions for analysis.

12
Objectives Of The PCA

• Captured the most important variable component
• The Composite Index
• Simplified the categorical factor
• Explained about the relationship between the
  variables

13
Categorical Factor
14
Analysis Flow Chart
15
Ranked Factor
16
Discriminate Analysis

• Significance
• R. A. Fisher (1936)

17
Dsicriminate Function
18
New Factor

• Significance motivation
• Improve the successful probability of the
• discriminate analysis result.

• Focus
• easy access to the information
• landmark

19
Erosion Factor (K)

• USDA-SCS (1978)

20
Things To Do

• Application for the PCA
• Find the new factor and other discriminate
  analysis
• Analysis results feasibility analysis study
• Write thesis

21
Thanks For Your Attention
22
Correlation Coefficient Matrix
23
The Number Of Principal Components Eigenvalue
24
Scree Plot
25
Rotated Component Matrix Formula
SAP1 0.17N0.76NE0.818E0.575ES
-0.065S-0.687SW-0.819W-0.433WN SAP2
-0.836N-0.396NE0.219E0.63ES
0.817S0.5SW-0.149W-0.719WN
26
T Test Table
27
Correlation Coefficient Table