國內目前事故分析相關研究，仍著重在公路或車輛上，對於鐵路與公路二大運輸系統之平面交會處的平交道甚少著墨。而平交道係鐵路事故最頻繁與死傷最嚴重之處，也是最需改善的項目。平交道事故之發生頻率與事故規模使其平交道具有高度風險，為避免事故發生所造成龐大的生命與財產損失，為平交道安全工作的重點。先進國家對平交道從事長期觀察與資料蒐集，每年進行平交道風險評估與危險平交道改善，但國內平交道風險之資料蒐集與分析仍十分缺乏。

本研究之宗旨是建立與測試平交道風險或事故之統計分析模式，一則可以具之預測平交道之事故發生機率或風險，一則可以說明平交道之風險因素。民國87年台灣省政府交通處出版一套四冊之平交道改善規劃，蒐集全省之平交道相關資料，包括84年至86年三年之平交道事故資料，與列車次數與公路每日交通量等重要曝光量資料。本研究利用上述資料，建立「平交道事故頻率模式」、「平交道事故衝擊模式」、與「平交道風險模式」，分別探討可能性指標-「單位時間之事故數」、衝擊性指標-「事故之等效死亡」、與風險指標-「單位時間之等效死亡」。

研究發現：事故頻率模式適合以卜瓦松迴歸分析，事故衝擊模式與事故風險模式適合負二項迴歸分析；影響平交道事故相關模式的因素中，以列車班次及平均每日交通量為最重要的風險因子。

The consequence and frequency of crossing are the most serious event of railway accidents, the advanced countries continue to be of concern despite an ever-increasing focus on improved design and practices, but little domestic research has examined the crossing risk.

The aim of this research is to evaluate the crossing risk and structure statistical models which are not only prediction of crossing accidents probability or risk, but explaining the risk factors. The data is from Crossing Improvement Plan by Taiwan Provincial Government in 1998, including three years accidents data from 1995 to 1997. The study establishes “Crossing Accident Frequency”, “Crossing Accident Impact”, and “Crossing risk” models to discuss respectively the Probability Index for accident numbers in unit time, the Impact Index for equivalent fatality of accidents, and the Risk Index for equivalent fatality in unit time.

This study finds the Crossing Frequency model is fit to Poisson regression; the Accident Impact and Accident Risk models are fit Negative Binomial regression.

The findings show promise for the methodological approach undertaken and provide important directions for future railroad research.

1.台灣省政府交通處，台灣地區鐵路平交道改善規劃(一)~(四)，民國87年。
2.立法院秘書處，「鐵路法修正案」，民國70年6月。
3.台灣鐵路管理局電務處，「電務規章彙編」，民國83年1月。
4.陳順宇，「統計學」，民國八十七年。
5.王文麟，「交通工程學-理論與實用」，民國八十一年。
6.許添本、李明聰，「地區性道路交道安全風險評估模式之建立」，道路交通安全評估與執法研討會，民國89年6月。
7.蔡明志，「風險管理在大眾運輸安全管理管制課題之發展應用」，運輸計畫季刊，第二十九卷第一期，181~212 頁，民國89年。
8.林豐福、周永暉，「台灣地區鐵路平交道事故之研究」，交通部運輸研究所，民國85年5月。
9.施鈞明，「應用系統模擬於平交道風險之分析」，國立成功大學交通管理科學系研究所碩士論文，民國93年1月。
10.楊佳學，「智慧型平交道安全監測系統安全評價之研究」，淡江大學運輸管理學系研究所碩士論文，民國92年6月。
11.鄭子良，「先進偵測系統對鐵路平交道安全之影響研究」，國立中央大學土木工程研究所碩士論文，民國91年7月。
12.張應輝，「台鐵營運安全風險之研究」，國立交通大學交通運輸研究所，民國90年6月。
13.駱思斌，「鐵路事故嚴重程度之研究」，國立成功大學交通管理科學系研究所碩士論文，民國90年6月。
14.陳志和，「都市地區肇事嚴重程度預測模式之研究」，國立成功大學交通管理科學系研究所碩士論文，民國88年6月。
15.戚培芳，「中山高速公路肇事分析模式之研究」，國立交通大學交通運輸研究所碩士論文，民國86年6月。
16.連仁宗，「卜瓦松模式於高速公路肇事預測之研究」，國立成功大學交通管理科學系研究所碩士論文，民國84年6月。
17.林豐福、張開國、吳熙仁，「應用風險管理於鐵路平交道改善之研究」，交通部運輸研究所，民國94年。
18.葉名山、周永暉，「台鐵行車安全問題掃描與剖析」，中華民國第一屆運輸安全研討會，民國83年11月。
19.台灣鐵路管理局網站，http://www.railway.gov.tw
20.交通部網站，http://www.motc.gov.tw
21.Frank, S. and Xiaoming, L., 2004, “Model for Evaluating Counter-Measures at Highway-Railway Grade Crossings,” Proceedings of the 84th Annual Meeting of the Transportation Research Board, Washington, D.C., U.S.A.
22.Gitelman, V. and Hakkert, A.S., 1997, “The Evaluation of Road-Rail Crossing Safety with Limited Accident Statistics,” Accident Analysis and Prevention. Vol. 29, No. 2, pp.171-179.
23.Greene, W., 1993, “Econometric Analysis,” Macmillan Publishing, New York.
24.James, R. L. and William, G. J., 1998, “Train Accident Reconstruction and FELA & Railroad Litigation Third Edition,” Lawyers & Judges Publishing Co.
25.Jutaek, O., Simon, W. and Doohee, N., 2004, “Accident Prediction Model for Railway-highway Interfaces,” Proceedings of the 84th Annual Meeting of the Transportation Research Board, Washington, D.C., U.S.A.
26.Office of Policy and Program Development, 2002, Reference Manual.
27.Richard, A. R. and Roy, E. L., 2004, “Examination of Highway-Rail Grade Crossing Collisions Over 10 Years in Seven Midwestern States,” Proceedings of the 84th Annual Meeting of the Transportation Research Board, Washington, D.C., U.S.A.
28.Ross, D. A. and Jodi, L. C., 2002, “An Alternative Accident Prediction Model for Highway-Rail Interfaces,” Accident Analysis and Prevention, Vol. 34, pp.31-42.
29.Sudhir, A. and Carl, D. M., 1998, “Level crossing safety on East Japan Railway Company Application of Probabilistic Risk Assessment Techniques,” Transportation, NO.25, p265-286.
30.Young-Jin P. and Frank, F. S., 2004, “Evaluating Factors Affecting Safety at Highway-Railway Grade Crossings,” Proceedings of the 84th Annual Meeting of the Transportation Research Board, Washington, D.C., U.S.A.
31.U.S.D.O.T., 1986, Railroad-Highway Grade Crossing Handbook—Second Edition,.
32.Davis Associates Limited, 2005, “Level Crossings Summary of Findings and Key Human Factors Issues,” Health and Safety Executive.
33.Brian, A., 2002, “Provision, Risk Assessment and Review of Level Crossings,” Railway Group Standard.
34.Brian, A., 2002, “Guidance on Provision, Risk Assessment and Review of Level Crossings,” Railway Group Standard.
35.Ashish, S., and Muni, S., 1990, “Regression Analysis-Theory, Methods, and Applications,” Springer.
36.Michael, W., 2004, “Development of a Program of Level Crossing Research to Improve Railway Safety in Great Britain,” Rail Safety and Standards Board.
37.Haibin, L., Rachel, A., Rosowsky, M., and Jery, R., 2005, “Negative Binomial Regression of Electric Power Outages in Hurricanes,” Journal of Infrastructure System, Vol. 11, No. 4, pp. 258-267.
38.Miaou, S. P., and Lum, H., 1993, “Modeling Vehicle Accidents and Highway Geometric Design Relationships,” Accident analysis and Prevention Vol 25, No. 6, pp. 689-709.
39.Miaou, S. P. and Hu, P. S., Wright, T., Rathi, A. K. and Davis, S. C., 1992, “Relationship Between Truck Accidents and Highway Geometric Design : A Poisson Regression Approach,” Transportation Research Record 1376, pp.10-18.
40.Miaou, S. P., 1994, The Relationship Between Truck Accidents and Geometric Design of Road Sections: Poisson versus Negative Binomial Regression,” Accident Analysis and Prevention, Vol. 26, No. 4, pp.471-482.
41.Jinsun, L., Doohee, N., and Dongjoo, P., 2005, “Analyzing the Relationship Between Grade Crossing Elements and Accidents,” Journal of the Eastern Asia Society for Transportation Studies, Vol. 6, pp. 3658-3668.
42.Lambert, D., 1992, “Zero-Inflated Poisson regression, with an application to defects in manufacturing,” Technometrics 34, pp. 1-13.
43.William, H. Greene, 1995, “LIMDEP Version 7.0 User’s Manual,” Econometric Software, Inc.
44.Vuong, Q. H., 1989, “Likelihood ratio tests for model selection and non-nested hypotheses,” Econometrica, Vol. 57, No. 2, pp. 307-333.
45.McCullagh, P., Nelder, J. A., 1989, “Generalized Linear Models,” Chapman and Hall.
46.Dimitris, K., Ioannis N., 2003, “Analysis of sports data by using bivariate Poisson models,” The Statistician, Vol. 52, Part 3, pp. 381-393.
47.Kocherlakota, S and Kocherlakota, K., 2001, “Regression in the Bivariate Poisson Distribution,” Commun, Statist- Theory Meth, 30, No. 5, pp. 815- 825.
48.Kocherlakota, S and Kocherlakota, K., 1992, “Bivariate Discrete Distributions,” New York: Dekker.
49.Johnson, N., Kotz, S. and Balakrishnan, N., 1997, “Discrete Multivariate Distributions,” New York: Wiley.
50.Wang K., Lee A.H., Yau K. K. W., and Carrivick P.J.W., 2003, “A bivariate zero-inflated Poisson regression model to analyze occupational injuries,” Accident Analysis and Prevention, Vol. 35, No. 4, pp. 625-629.
51.Robert C.J. and Rainer W., 1993, “Two aspects of labor mobility: A bivariate Poisson regression approach,” Empirical Economics, Vol. 18, No.3, pp. 543-556.
52.Lee A.H., Wang K., Yau K.K.W., Carrivick P.J.W. and Stevenson M.R., 2005, “Modelling bivariate count series with excess zeros,” Mathematical Biosciences, 196, pp.226-237.
53.Ho L.L., and Singer J.M., 2001, “Generalized least squares methods for bivariate poisson regression,” Commun. Statist.-Theory Meth., 30(2), pp. 263-277.
54.Selvin, S., 1995, “Practical biostatistical methods,” Duxbury Press.