臺灣博碩士論文加值系統

交通違規為影響道路安全及造成交通事故的主要原因之一，然而違規資料龐大且複雜，如何從中找出有用的資訊、建立合適的分析模型、運用互動式的視覺化工具，以了解違規資料的相關性與趨勢及違規駕駛行為。本論文將建立一個違規資料分析模型與使用者視覺化互動平台，運用真實違規資料進行分析，讓使用者透過互動式的人機介面，觀察到違規資料的趨勢變化及違規特性。首先，我們針對違規資料瞭解其欄位定義與資料型態，將原始資料進行資料清理，排除格式錯誤、欄位值超出範圍等異常情形之資料。接著，運用違規資料進行分析，再將資料進行分類，去除對分類不具影響的屬性，並利用具關鍵性影響的屬性，使用隨機森林演算法分類，進行空間分析、時間分析、行為分析等找出重大違規特徵值。找出不同年齡層違規差異，汽車和機車易違規條款。並根據台灣北部、中部、南部等不同區域進行比較，找出地區違規差異。分析結果進行視覺化呈現，可選擇並根據不同時間、地點、年齡、性別等看出違規數量或違規條款差別等變化。使用Highcharts與C#等開發工具來顯示分析結果，如熱區圖、樞紐分析圖等強調資料明顯差距。最後根據分析結果，使用者可透過該平台找出重大違規事件，可看出各地區易違規區域，提供政府或警察單位做為未來執法和道路安全建設之依據。以此方式改善道路安全以及有效降低肇事發生。

Traffic violation is one of the major reasons of road safety preventions and traffic accidents, but the violation data is too large and complex. Thus, extracting usable information, building analysis models, and using interactive tools are viral to understand the relevance, trends, and driving behaviors from the traffic violation data. In this paper, we will establish an analysis model that suitable for traffic violation data and visualization platform for users by using historical traffic violation data of Taiwan in recent four years. First, we will focus on using the traffic violation data to understand the definitions and data types of violation attributes in order to exclude the data abnormality, including formatting errors and over-range. Second, we remove the attributes that have no effect on the classification through the violation data analysis and classification which were adopted by using the random forest scheme on the H2O platform for those attributes with the most critical impacts. We also perform the spatial, temporal, and behavior analysis to find the features of serious violations. Analysis of violation discrepancy is performed for variety of time, space, and ages, and different traffic penalty rules for cars and motorcycles. By comparing the information from Taiwan's northern, central, and southern cities, we also investigate in the regional differences in violations. After the analysis, we do the visualization to show the variety on the basis of the number of violations and violation rules. We use Highcharts and C# development tools to show the analysis results, and emphasize the obvious differences of data such as Heatmap and PivotTable report. Finally, users can use the platform to find critical traffic violations based on the analysis results. For examples, government officers or policeman with information can strengthen the law enforcement and road construction of road safety in the future.

誌謝
摘要
Abstract
Table of Contents
List of Figures
List of Tables
Chapter 1 Introduction
1.1 Background
1.2 Motivation
1.3 Overview of Research
1.4 Thesis Organization
Chapter 2 Related Work
2.1 Traffic Violation
2.2 Supervised Learning
2.3 Visualization System
Chapter 3 Traffic Violation Analysis and Visualization
3.1 System Overview
3.2 Data Collection
3.3 Data Preprocessing
3.3.1 Data Standardization
3.3.2 Data Transformation
3.3.3 Descriptive Statistics
3.4 Serious Violation Classification
3.4.1 Analyze Model Architecture
3.4.2 Random Forest
Chapter 4 System Implementation and Prototype
4.1 System Implementation
4.2 System Prototype
Chapter 5 Experimental Results
5.1 Experimental Environment
5.2 Appropriate Parameters Testing
5.3 Three Cities Differences
Chapter 6 Conclusions
References

[1]E. Abdullah and A. Emam, "Traffic Accidents Analyzer Using Big Data," in Proceedings of the International Conference on Computational Science and Computational Intelligence (CSCI), Las Vegas, NV, USA, pp. 392-397, Mar. 2016.
[2]A. Addi, A. Tarik, and G. Fatima, "An Approach Based on Association Rules Mining to Improve Road Safety in Morocco," in Proceedings of the International Conference on Information Technology for Organizations Development (IT4OD), Fez, Morocco, pp. 1-6, May. 2016.
[3]J. Ahn, E. Ko, and E. Y. Kim, "Highway Traffic Flow Prediction Using Support Vector Regression and Bayesian Classifier," in Proceedings of the International Conference on Big Data and Smart Computing (BigComp), Hong Kong, China, pp. 239-244, Jan. 2016.
[4]F. Alencar, C. M. Filho, and D. G. d. Silva, "Pedestrian Classification Using K-means and Random Decision Forests," in Proceedings of the Joint Conference on Robotics: SBR-LARS Robotics Symposium and Robot Control (SBR LARS Robot Control), Sao Carlos, Brazil, pp. 103-108, Oct. 2014.
[5]D. Anggraini, W. Siswantoko, and D. Henriyan, "Design and Implementation of System Prediction and Traffic Conditions Visualization in Two Dimensional Map (Case Study: Bandung City)," in Proceedings of the 6th International Conference on System Engineering and Technology (ICSET), Bandung, Indonesia, pp. 87-91, Feb. 2016.
[6]F. Babič and K. Zuskáčová, "Descriptive and Predictive Mining on Road Accidents Data," in Proceedings of the 14th IEEE International Symposium on Applied Machine Intelligence and Informatics (SAMI), Herlany, Slovakia, pp. 87-92, Jan. 2016.
[7]H. Cai, D. Zhu, and L. Yan, "Using Multi-regression to Analyze and Predict Road Traffic Safety Level in China," in Proceedings of the Transportation Information and Safety (ICTIS), Wuhan, China, pp. 363-369, Jun. 2015.
[8]P. Cruz, A. Cruz, and P. Machado, "Contiguous Animated Edge-Based Cartograms for Traffic Visualization," IEEE Computer Graphics and Applications, vol. 35, pp. 76-83, Sep. 2015.
[9]P. Cruz and P. Machado, "Pulsing Blood Vessels: A Figurative Approach to Traffic Visualization," IEEE Computer Graphics and Applications, vol. 36, pp. 16-21, Mar. 2016.
[10]R. ElHakim, M. Abdelwahab, and A. Eldesokey, "Traffisense: A smart integrated visual sensing system for traffic monitoring," in Proceedings of the SAI Intelligent Systems Conference (IntelliSys), London, UK, pp. 418-426, Nov. 2015.
[11]V. Flores, M. Mata, and J. Fernandez, "A Multi-agent, In-vehicle Database Recorder System for Supporting Traffic Hotspots Detection, Geographical Representation and Analysis View Document," in Proceedings of the IEEE International Conference on Information Fusion (FUSION), Salamanca, Spain, pp. 418-426, Jul. 2014.
[12]R. Grodi, D. B. Rawat, and F. Rios-Gutierrez, "Smart parking: Parking occupancy monitoring and visualization system for smart cities," in Proceedings of SoutheastCon, Norfolk, VA, USA, pp. 1-5, 2016.
[13]A. Imawan and J. Kwon, "A Timeline Visualization System for Road Traffic Big Data," in Proceedings of the IEEE International Conference on Big Data (Big Data), Santa Clara, CA, USA, pp. 2928-2929, Nov. 2015.
[14]A. Jahangiri, H. A. Rakha, and T. A. Dingus, "Adopting Machine Learning Methods to Predict Red-light Running Violations," in Proceedings of the 18th IEEE International Conference on intelligent Transportation Systems (ITSC), Las Palmas, Spain, pp. 650-655, Nov. 2015.
[15]X. Ji, Z. Wei, and L. Huang, "Violation Vehicle Automated Snap and Road Congestion Detection," in Proceedings of the 4th International Conference on Cloud Computing and Intelligence Systems, Beijing, China, pp. 16-20, Aug. 2016.
[16]N. Joshi, A. K. Tripathy, and S. Sawant, "Near real time vehicle tracking using GIS," in Proceedings of the International Conference on Technologies for Sustainable Development (ICTSD), Mumbai, India, pp. 1-6, Feb. 2015.
[17]T. Landesberger, D. W. Fellner, and R. A. Ruddle, "Visualization System Requirements for Data Processing Pipeline Design and Optimization," IEEE Transactions on Visualization and Computer Graphics, vol. 23, pp. 2028 - 2041, Aug. 2016
[18]H. Liu, A. K. Tk, and J. P. Thomas, "Cleaning Framework for BigData: An Interactive Approach for Data Cleaning," in Proceedings of the IEEE Second International Conference on Big Data Computing Service and Applications (BigDataService), Oxford, UK, pp. 174-181, Apr. 2016.
[19]T. Lu, Z. Dunyao, and Y. Lixin, "The Traffic Accident Hotspot Prediction: Based on the Logistic Regression Method," in Proceedings of the International Conference on Transportation Information and Safety (ICTIS), Wuhan, China, pp. 107-110, Jun. 2015.
[20]G. Ou, Y. Gao, and Y. Liu, "Real-Time Vehicular Traffic Violation Detection in Traffic Monitoring Stream," in Proceedings of the IEEE/WIC/ACM International Conferences on Web Intelligence and Intelligent Agent Technology (WI-IAT), Macau, China, pp. 15-19, Dec. 2012.
[21]R. Ranjith, J. J. Athanesious, and V. Vaidehi, "Anomaly Detection Using DBSCAN Clustering Technique for Traffic Video Surveillance," in Proceedings of the Seventh International Conference on Advanced Computing (ICoAC), Chennai, India, pp. 1-6, Dec. 2015.
[22]N. Tanaka and T. Noguchi, "Vehicle visualization system on the iOS devices using VANETs," in Proceedings of 2016 Eighth International Conference on Ubiquitous and Future Networks (ICUFN), Vienna, Austria, pp. 154-159, Jul. 2016.
[23]W. Tong, P. Cherian, and J. Liu, "Statistical Analysis of DMV Crash Data," in Proceedings of the IEEE International Congress on Systems and Information Engineering Design Symposium (SIEDS), Charlottesville, VA, USA, pp. 113-117, Apr. 2016.
[24]E. Trudel, C. Yang, and Y. Liu, "thr Data-driven Modeling Method for Analyzing Grade Crossing Safety," in Proceedings of the 20th IEEE International Conference on Computer Supported Cooperative Work in Design (CSCWD), Nanchang, China, pp. 145-151, May 2016.
[25]D. Yokoyama and M. Toyoda, "Do Drivers' Behaviors Reflect Their Past Driving Histories? - Large Scale Examination of Vehicle Recorder Data," in Proceedings of the IEEE International Conference on Big Data (BigData Congress), San Francisco, CA, USA, pp. 361-368, Jul. 2016.
[26]D. Yokoyama and M. Toyoda, "A Large Scale Examination of Vehicle Recorder Data to Understand Relationship Between Drivers' Behaviors and Their Past Driving Histories," in Proceedings of the IEEE International Conference on Big Data (Big Data), Santa Clara, CA, USA, pp. 2877-2879, Oct. 2015.
[27]H2O, http://www.h2o.ai/
[28]NCHC, https://www.nchc.org.tw/tw/