結合資料插補概念與資料探勘技術於旅客再住宿意願之分析

Data Mining Techniques with Data Imputation in Analyzing The Re-accommodation Willness of Visitors

楊昇勳(Sheng-Hsun Yang)

指導教授 ： 白炳豐

國立暨南國際大學/管理學院/資訊管理學系/碩士(2012年)

摘要

近年來隨著社會經濟的成長，台灣民眾外出旅遊的人數變多，逐漸的形成國民旅遊的風氣。伴隨著國內民眾休閒旅遊的盛行，旅遊特性也相對的轉為多元變化，促進了民宿市場的蓬勃發展。民宿聚集地座落在觀光熱門景點，而住宿旅客則為促進當地觀光景點產業的重要因子。面對住宿旅客需求資訊的產生，民宿業者也體認到個別一家民宿業者所能掌握的住宿旅客的資訊是有限的，但過多的資訊卻容易對人造成決策混亂。因此，藉由資訊探勘技術的應用建造規則庫，相信能夠幫助各領域做出最佳的分析與決策。
在本研究中首先將利用遺漏值處理方法，針對民宿產品意象、品牌性格與再宿意願的問卷填答，進行遺漏值資料的改善，並結合特徵擷取的技術互相比較：首先利用遺漏值處理方法，如刪除法和多重插補法進行遺漏值處理，接著各別使用差分演算法，針對資料集進行特徵擷取選出關鍵特徵，比較不同的遺漏值處裡方法擷取出的特徵差異。最後利用決策樹C4.5產生最佳的準確率及樹狀結構規則，以期建立一旅客再住宿意願規則庫，讓民宿業者在滿足住宿旅客需求及提升遊客再次住宿意願能有所依據。

關鍵字

遺漏值 ； MICE ； 差分演算法 ； 決策樹C4.5

並列摘要

With the socio-economic growth in recent years, the people of Taiwan go out to the number of tourists become more, the formation of the national tourism culture. Along with the prevalence of domestic public leisure and tourism, and tourism characteristics relative to the multiple changes and promote the vigorous development of the Bed and Break market.
B & B fast gathering place is located in the popular tourist attractions and accommodation visitors were an important factor to promote local attractions industry. Guests as demand for information generated, guesthouse owners recognize the individual a guesthouse owners can master the guests as information is limited. It’s easier on the human decision-making confusion.
This study will take advantage of the approach of missing values, missing values information to improve the questionnaires. For B & B product image, brand personality and repurchase intention, and the combination of feature extraction techniques to compare. First, the missing values handling methods, such as delete method and multiple imputation method for missing values and then the individual is added to the differential algorithm, feature extraction, and data characteristics to elect key features. Secondary, compared the feature extraction from the differential evolution. Finally, in the accuracy of the C4.5 decision tree and the tree structure rules. In order to establish a passenger re-accommodation rule base, so the guesthouse owners to meet the needs of guests as again stay will be able basis.

並列關鍵字

Missing value ； MICE ； DE ； C4.5

參考文獻

A.Qing, Xu, X., & Gan, Y. B. (2005). Anisotropy of composite materials with inclusion with orientation preference. IEEE Antennas and Propagation Society, 53(2), 737-744.

Aaker, D. A. (1996). Building strong brands. New York: The Free Press.

Aaker, D. A. (1997). Managing brand equity. New York: The Free Press.

Aaker, J. L., Benet-Martinez, V., & Garolera, J. (2001). Consumption symbols as carriers of culture: A study of Japanese and Spanish brand bersonality constructs. Journal of Personality and Social Psychology, 81(3), 492-508.

Abbass, H. A., Sarker, R., & C. Newton. (2001). PDE: a Pareto-frontier differential evolution approach for multi-objective optimization problems. Paper presented at the Proceedings of IEEE Congress on Evolutionary Computation, NJ.