交通意外事故每年都造成嚴重傷亡，因此，為了有效降低交通事故傷亡，針對交通意外事故加以分析非常重要。本研究採用隨機森林演算法針對2015-2016年花蓮縣交通警察隊交通事故去識別化資料進行資料探勘。主要內容可以分成兩部分，第一部分針對主要肇因為「未注意車前狀況」的車禍個案，探討道路環境因子對車禍嚴重程度的影響；第二部分針對所有駕駛人，找出影響車禍嚴重程度的因子。結果顯示事故型態、事故位置為影響事故嚴重程度的道路環境因子，發生在交岔路口內或交岔路口附近的事故最為嚴重。腳踏車、機車駕駛人，以及行人在發生事故時，傷亡較為嚴重。另外事故發生月份及星期，也是影響事故嚴重程度的因子。

Traffic accident has been one of major causes of death in Taiwan. Thus, it is important to effectively analysis the traffic accident patterns in order to reduce traffic accidents. This study collects the 2015-2016 de-identified traffic accident data from Traffic Police Brigade, Hualien County Police Bureau. We investigate traffic accident data by random forest algorithm. The whole study is divided into two parts. The first part focuses on environmental factors for inattentive accidents which will result in more severe damage. The second part is to realize at-fault driver factors which have great influence on causing severe accident or the property damage. The results suggest that accident type and accident position are the dominant environmental factors for severe traffic accidents. Accidents happened at intersection or nearby intersection are the most severe. Bicyclists, motorcyclists, and pedestrians caused more severe traffic accidents. Furthermore, accident happened month and weekday are also the significant factors for accident severity.

Chapter 1 Introduction 1
1.1 Background 1
1.2 Motivation 4
1.3 Research Purpose 5
1.4 Study Process 7
Chapter 2 Literature Review 9
2.1 Machine Learning Methods in Prior Study 9
2.1.1 Regression Model 9
2.1.2 Classification Model 11
2.2 Tree-based Model 17
2.2.1 Feature Selection 17
2.2.2 Branch Pruning 18
2.2.3 Ensemble Learning 18
2.3 Summary 20
Chapter 3 Data and Methodology 21
3.1 Data Source and Tasks 21
3.1.1 Data Form 22
3.1.2 Data Mining Tasks 24
3.2 Methodology 25
3.2.1 Machine Learning 25
3.2.2 Random Forest 25
3.2.3 Feature Importance 26
3.2.4 Cross Validation 26
Chapter 4 Experimental Analysis 29
4.1 The Public-facility Relative Factors 29
4.1.1 Inattentive Accident 30
4.1.2 Non- inattentive Accident 37
4.2 At-fault Driver Factor 41
Chapter 5 Conclusion and Recommendation 51
5.1 Conclusion 51
5.2 Recommendation 52
Appendix A. Sample of Table I and Table II 55
Appendix B. Causes of Traffic Accident 57
Appendix C. Parallel English-Chinese Texts 58
References 60

CHANG, Li-Yen; CHIEN, Jui-Tseng. Analysis of driver injury severity in truck-involved accidents using a non-parametric classification tree model. Safety science, 2013, 51.1: 17-22.
CHANG, Li-Yen; WANG, Hsiu-Wen. Analysis of traffic injury severity: An application of non-parametric classification tree techniques. Accident Analysis & Prevention, 2006, 38.5: 1019-1027.
DELEN, Dursun; SHARDA, Ramesh; BESSONOV, Max. Identifying significant predictors of injury severity in traffic accidents using a series of artificial neural networks. Accident Analysis & Prevention, 2006, 38.3: 434-444.
HALEEM, Kirolos; GAN, Albert. Effect of driver’s age and side of impact on crash severity along urban freeways: A mixed logit approach. Journal of safety research, 2013, 46: 67-76.
HARB, Rami, et al. Exploring precrash maneuvers using classification trees and random forests. Accident Analysis & Prevention, 2009, 41.1: 98-107.
KAPLAN, Sigal; PRATO, Carlo Giacomo. Risk factors associated with bus accident severity in the United States: A generalized ordered logit model. Journal of safety research, 2012, 43.3: 171-180.
KOCKELMAN, Kara Maria; KWEON, Young-Jun. Driver injury severity: an application of ordered probit models. Accident Analysis & Prevention, 2002, 34.3: 313-321.
MONTELLA, Alfonso, et al. Analysis of powered two-wheeler crashes in Italy by classification trees and rules discovery. Accident Analysis & Prevention, 2012, 49: 58-72.
LI, Zhibin, et al. Using support vector machine models for crash injury severity analysis. Accident Analysis & Prevention, 2012, 45: 478-486.
YAN, Xuedong; RADWAN, Essam; ABDEL-ATY, Mohamed. Characteristics of rear-end accidents at signalized intersections using multiple logistic regression model. Accident Analysis & Prevention, 2005, 37.6: 983-995.
DE OÑA, Juan; MUJALLI, Randa Oqab; CALVO, Francisco J. Analysis of traffic accident injury severity on Spanish rural highways using Bayesian networks. Accident Analysis & Prevention, 2011, 43.1: 402-411.
SAVOLAINEN, Peter; MANNERING, Fred. Probabilistic models of motorcyclists’ injury severities in single-and multi-vehicle crashes. Accident Analysis & Prevention, 2007, 39.5: 955-963.
FRIEDMAN, Jerome; HASTIE, Trevor; TIBSHIRANI, Robert. The elements of statistical learning. New York: Springer series in statistics, 2001.
GRUS, Joel. Data science from scratch: first principles with python. " O'Reilly Media, Inc.", 2015.
OZDEMIR, Sinan. Principles of Data Science. Packt Publishing Ltd, 2016.
UNPINGCO, José. Python for probability, statistics, and machine learning. Springer International Publishing, 2016.