Abstract—This study used GIS to determine 19 topographic
indexes, four geologic indexes, and two rainfall data derived
indexes of the Laonun River Basin, in southern Taiwan. The
four topographic factors, including the effective area of basin,
elongation ratio of basin, relief energy, and relief volume, were
selected among the 19 topographic indexes using the SPSS
multi-variable statistical analysis and the principal components
analysis. The four topographic factors combine the four geologic
factors and two rainfall factors and were selected for estimating
debris flow prone creeks. The ten factors were further applied
by the Fisher's Discriminant Analysis and Logistic Regression
Analysis to evaluate the potentials of debris flow prone creeks in
the basin. There were 13 sub-basins initiated debris flows and 41
non-debris flow sub-basins when Typhoon Morakot hit Taiwan
in 2009. The validated results show that the correctness of
Fisher’s model for the samples is 81.48 % and 92.6 % via the
Logistic Regression model. Both models showed acceptable
accuracy, and the Logistic model had better accuracy herein.
The Logistic Regression Analysis was adapted to evaluate the
potential of debris flow sub-basins to assist in developing risk
management in the basin.
Index Terms—Debris flow, GIS, principal components
analysis, linear regression analysis.
Chien-Yuan Chen and Zhao-Jun Chen are with the Civil and Water
Resources Engineering Department, National Chiayi University, Chiayi City
60004, Taiwan R.O.C. (e-mail: chienyuc@mail.ncyu.edu.tw,
tainan946012@yahoo.com.tw).
Ho-Wen Chen is with the Department of Environmental Science and
Engineering, Tunghai University, Taichung City 40704, Taiwan R.O.C.
(e-mail: hwchen@thu.edu.tw).
Cite: Chien-Yuan Chen, Ho-Wen Chen, and Zhao-Jun Chen, "Determination of Topographic Factors to Initiate Debris Flow Using Statistical Analysis," International Journal of Machine Learning and Computing vol. 4, no. 6, pp. 547-552, 2014.