透過您的圖書館登入
IP:34.205.142.9
  • 學位論文

應用子平學於交通事故分析及預測之研究

Traffic Accident Analysis and Prediction—an Application of Zih-Ping

指導教授 : 張堂賢
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


安全與效率是交通科學追求的目標。回顧相關研究,發現目前的研究所使用的參數,不外乎從「人」、「車」、「路」,即駕駛行為、車輛狀況與道路環境等著手研究。中國傳統方法之參數是觀察大自然而來,包括了陰陽、五行、八卦、天干、地支…等。起初為描述大自然現象,然因中國人講求天人合一、順天應人,經模式轉移後便可將此理論應用於預測「人」的禍福吉凶。相較於現有相關之研究方法,必須發生非常多事故後方能進行肇事分析,傳統方法為較人道之作法。 本研究嘗試以中國傳統方法所使用的參數,對交通事故進行分析及預測,所採用的傳統方法為子平八字,期許能從肇事資料中驗證現有易肇事之八字結構,並使用窮舉法與資料探勘的方法歸納出易肇事之八字結構。 本研究主要分成三部份:(1)驗證現有易肇事八字結構,(2)以八字基本理論為基礎,窮舉出各干支間關係之排列組合,找出發生機率較高之八字結構,(3)以現有肇事資料為輸入資料,在沒有八字理論的基礎下,使用資料探勘之決策樹演算法找出事故類別之判斷準則。 研究指出,驗證之九條現有易肇事八字結構中,僅有一條樣本機率大於母體機率,但差異僅0.125%。窮舉八字各干支會合刑沖等關係後,發現現有肇事資料中,出生月干與大運天干若有「剋」、「五合」關係,以及肇事年干與命局地支(出生年支、月支、日支)有「生」關係者,其發生比例較母體發生機率高5%以上,最高者接近8%。另外,分別輸入八字資料為干支型態與五行型態,使用決策樹演算法得到之事故類別判斷準則,準確率分別為80.0%與76.6%。

並列摘要


Safety and efficiency are the main objectives in traffic science. Upon all related literatures, we discover the parameters which have been used in current researches are focus on the road user, vehicle and road. The Chinese traditional methods includes utilize Yin Yang, Five Elements, Eight Trigrams, Heavenly Stems, and Earthly Branches. Initially, people observed the natural phenomenon, and then transfer the theory concluded by the natural observation as a method to predict the fortune because Chinese believed that human and natural live together, and cannot live individually. Current research methods should be analyzed with many car accidents; therefore, traditional methods are more humanly compared with current researches methods. In this research, we try to carry out the traffic accident analysis and forecasts via utilizing the Zin-Ping Eight Words to prove the relation between the existing traffic accident data and the Eight Words structures of victims and conclude some Eight Words structures with higher probability to confront the traffic accidents by the exhaustive method and data mining. The result of this research, a analysis of nine Eight Words Structures with higher probability to confront the accidents, only one comparison result is higher than the population probability and the difference is only 0.125%. Base on reciting the relation with the Heavenly Stems and Earthly Branches by existing traffic accident data, the results indicate that the traffic accident probability will be 5% over population probability, even close to 8%, if occurs the specific relation between the Heavenly Stems and Earthly Branches. The specific relations are given as below. The relation between Birth month Heavenly Stream and decade Heavenly Stream are “Ke” and “wehe”. The relation between Incident time year Heavenly Stream and zodiac fate Earthly Branches is “sheng”. Moreover, the results of Heavenly Stream and Earthy Branch analysis also indicate that the accuracy of the judge rules are 80.0% and 76.6% respectively through decision tree algorithms.

參考文獻


28. Al-Ghamde, A. S., “Using logistic regression to estimate the influence of accident factors on accident severity”, Accident Analysis and Prevention, 34(6), 2002, 729-741.
30. Fayyad, U. M., “Data mining and knowledge discovery: making sense out of data”, IEEE Expert, Vol. 11, No. 5,1996, pp.20-25.
31. Fayyad, U. M., G.. Piatetsky-shapiro, and P. Smyth, “From Data Mining to Knowledge Discovery in Databases”, AI Magazine,1996, pp.37-54.
32. U. Fayyad, G. Piatetsky-Shapiro and P. Symth ,”Overview of Data Mining and Knowledge Discovery, Knowledge Discovery and Data Mining”, AAAI press, 1996, pp. 1-36.
33. Karlaftis, M. G., Golias, I., “Effects of road geometry and traffic volumes on rural roadway accident rates”, Accident Analysis and Prevention, 24(3),2002, pp.357-365.

延伸閱讀