利用電子票証資料推估公車旅運起迄之研究

Abstract

大眾運輸旅客之旅次起迄型態分析可提供旅客在特定區域內之起訖數量、搭乘路線、旅行方向、路徑選擇、出發時間、旅次長度與轉乘等資訊。該資訊可作為運具路線之整併、班表調整、路網設計、費率訂定及政策之制定等重大的參考依據。因此如何獲得大眾運輸旅運起迄量為管理者所面臨的一項問題。 隨著科技的發展，電子票證收費系統已經廣泛的運用於大眾運輸系統。其電子票證交易紀錄含有乘客起迄之特性，可藉此資訊來獲得搭乘者之旅次起訖等資訊。因此透過分析大量的電子票證數據取得搭乘者之起迄已是刻不容緩。 然而受限於台灣部分地區公車之段次計費特性，在進行公車旅運起迄之推估時往往會有許多資料不足之狀況。因此本研究之目的在於設計一套方法來推估台灣段次計費公車之起迄。本研究透過乘客乘公車之刷卡站位、搭乘公車路線之方向以及其刷卡方式等資料，來設計各種乘客可能搭乘之情境來推估起迄點，並計算模式之誤差率。最後經由計算本研究模式之誤差率為49%，其誤差率比Gao(2015)之誤差率(76%)來的低。

關鍵詞：起迄點推估、資料探勘、電子票證

The analysis of origin and destination patterns of public passengers can provide the information of passenger’s origins and destinations, travel direction, route choice, departure time, trip distance and transfer mode. These information can be utilized to route integration, schedule adjustment, network design and making policies. Hence, it is important for administrators to obtains passenger’s origin and destinations. With the development of technology, electronic payment systems are being used more and more by public transit systems. The transactions records of electronics payment data contain the properties of passenger’s origins and destinations. These transactions records can be utilized to obtain passenger’s origins and destinations. Therefore, it is really necessary to analyze electronic payment data. Due to charging way of Taiwan bus, the electronic payment data might contain lots of fragmented data, such as the data lack of boarding or alighting points. Hence the objective of the research is to design a method to estimate the origins and destinations of passengers taking Taiwan bus. The research design the situations of passengers who take bus to estimate origins and destinations by utilizing the data of bus stop, direction and the way of charging. Finally, the error rate of the method designed in this research is 49% which is lower than the error rate(76%) of Gao(2015).

Keywords：O-D estimation、Data mining、Electronic Payment