結合隨機森林與複合態度模式探討國道客運創新服務發展與行銷策略之研究 – 以台北宜蘭線為例

Applying Random Forest and PIA Attitude Model to Explore New Service Development and Marketing Strategies for Inter-city Transit - Using Yilan as Case Study

蔡子彥(Tzu-Yen Tsai)

指導教授 ： 許超澤

淡江大學/商管學院/運輸管理學系碩士班/碩士(2016年)

https://doi.org/10.6846/TKU.2016.00975

摘要

在全球永續運輸思潮推動下，政府積極鼓勵民眾使用大眾運輸工具，但國人普遍偏好在國道使用私人運具，導致國道客運使用率遠低於其他大眾運輸。因此，如何提出適合且有效的運輸行銷策略，提升民眾搭乘意願，並減輕對環境的衝擊影響，已成為重要的研究課題。傳統上，大眾運輸行銷多以提升服務品質或降低票價來吸引民眾搭乘。而服務品質研究大多著眼於發掘並改善旅客不滿意的服務，雖可保有舊有客源，但對於拓展客源的效果有限。研究顯示，有效的大眾運輸行銷需要探索民眾對於大眾運輸的態度及潛在需求，並導入創新服務思維以滿足消費者期待，針對各族群提出不同的行銷策略。自80年代起，學者陸續提出不同的發展創新服務操作流程，但往往流於概念式的架構，缺乏可行的操作方法。因此，本研究彙整學者所提出之各類型創新服務發展架構，提出一個結合創新服務與運輸行銷的基本研究架構。
　　本研究建構以感知-意向-適應模型為基礎的態度問卷，透過隨機森林法進行市場區隔，探討不同族群對於國道客運之態度以及潛在需求，並據以建立分群預測模型。其後，與一般民眾進行半開放式問卷訪談，問卷結果會根據預測模型進行分類，將受訪者分配至相對應族群，探討各族群旅客的期望服務。再與國道客運業者討論各族群期望新服務的可行性，針對各族群擬定可行的行銷策略，並探討業者與民眾之間對於期望服務的差異及原因。
　　研究結果顯示，完整且有效的行銷策略需結合公、私部門共同投入，建立短、中及長期的整合計畫。短期行銷策略應著重於各族群皆期望且業者認為可立即提供同時可行性也較高的服務，如提供觀光景點一日券、景點直達車以及提供整合性的國道客運與觀光資源應用軟體等。中期行銷策略可依照業者所欲行銷的目標族群，提供該族群期望的新服務;同時，透過公部門的推動，逐步開放電子票證於國道客運上使用。長期的行銷策略則是以達成政府所提倡空間、時間、資訊及服務無縫的概念為目標，透過與公部門的合作，逐步建設相關的完整軟硬體基礎設施，提供更為便利的運輸環境。

關鍵字

創新服務發展 ； 行銷 ； 感知-意向-適應模型 ； 隨機森林

並列摘要

Due to the global trend of sustainable transportation, the government proactively encourages general public to use transit. However, most of people still prefer to use private vehicles on highway which significantly affect the ridership of inter-city transit. Therefore, it is important to develop suitable and efficient marketing strategies to increase ridership and mitigate the environmental impacts. Traditionally, the marketing strategies of transit mainly focus on reducing the fare or providing proper service quality to attract passengers. Using service quality research to explore unsatisfied services and then adjust them could increase the satisfaction level but it is difficult to attract new passengers. Study shows that to understand public attitude toward transit and latent demand is critical to develop effective strategies. Providing a variety of new services for each target groups to fulfill their expectation through market segmentation is the best way to attract more passengers. Starting from 1980s, scholars have been proposing many types of new service development process; however, they tend to provide a conceptual system lacking operational framework. Therefore, we combine previous research results to develop a simple operational framework for integrating new service development and transit marketing in this research.
　　This research first develops an attitude questionnaire based on Perception-Intention-Adaptive model and then uses random forest to conduct market segmentation in order to explore public attitude and latent demand for each group. Also, a prediction model is built based on the collected data. Second, we use the same questionnaire with open-ended questions to collect expected new services from public and use the prediction model to assign expectation to each target groups.Finally, the feasibility of expected new services is screened by the industrial experts and the differences between public and industrial experts are also discussed.
　　The study results show that integrating the efforts of government and transit industry to develop a series of comprehensive plans can build effective marketing strategies. Short term marketing strategies should focus on providing services that fulfills the expectation of all groups and also can be provided immediately by the industry, such as express services to each scenic site and full function integrated smartphone application. The midterm marketing strategies are to provide new services to targeted groups and allow the use of electronic payment. The long term marketing strategies are to achieve the goal of seamless transportation which is advocated by government. The cooperation of the government and the industry can provide an attractive and convenient transportation system.

並列關鍵字

New Service Development ； Marketing ； Perception-Intention-Adaptive ； Random Forest

參考文獻

1. 巫天虹（2013），以兩階段分類法建構信用卡授信決策模型的實務評估，淡江大學統計學系應用統計學碩士班碩士論文。

2. 李靜芳、陳耀茂、林寒琦、鍾健平（2012），企劃國道客運業新型服務之實證研究，明道學術論壇，6(4)，3-21。

4. 林伯勳（2013），使用隨機森林演算法做煙火偵測，國立暨南大學資訊工程學系研究所碩士論文。

6. 周建張、藍雅莉、莊麗秋、吳佩穎（2005），國光客運與統聯客運服務品質之比較，都市交通季刊，20(1)，33-44。

12. 陳明鴻（2012），針對「創新服務概念發展」之設計程序，國立成功大學工業設計學系碩博士班。