摘要
競爭是當今社會最普遍的現象，在生活中，我們都會有意無意的被卷入無數場競 爭當中。而企業之間的競爭更是異常的激烈，這種情況出現在任何產業當中。對 於每個產業來説，他們的業務都逐漸趨近於飽和，如何不斷增加收入和盈利能力 是所有企業都會面臨的重要問題。在很多文獻研究中可以發現吸引新客戶需要的 付出的成本遠高於保留現有客戶的成本。因此如何保留現有客戶是當今較受關注 的問題。因為成功的保留住現有客戶就有大幅的機會為公司在眾多競爭中取得優 勢。不過由於客戶流失是不可避免的課題，那麽問題就轉換成尋找客戶的購買模 式、內容以及其是如何影響客戶流失。購買的內容以及購買的模式代表了消費者 和超商之間的關係，超商在消費者心中的定位是什麼，這個關係會決定消費者會 與超商維持多久的關係，而購買某種類別型態的關係又會存活多久。在零售業大 家都會面臨一個困難點，這個客戶如果沒有來，他是否還存活著?還是已經流失? 因為流失這件事情不太容易定義，因此本篇論文使用分析的情境來定義流失，並 且將這些生存數據放入生存分析模型中，估計購買模式以及內容與生存時間的關 係，在本篇中我們應用一間日本連鎖品牌的超商作为研究對象。

ABSTRACT
Competition is the most common phenomenon in today's society. In our life, we will all be involved in countless competitions intentionally or unintentionally. And the competition between enterprises is extremely fierce, which occurs in any industry. For each industry, their business is gradually approaching saturation, and how to continuously increase revenue and profitability is an important issue that all companies will face. In many literature studies, it can be found that the cost of attracting new customers is much higher than the cost of retaining existing customers. Therefore, how to retain existing customers is a more concerned issue today. Because the successful retention of existing customers has a substantial opportunity for the company to gain an advantage in many competitions. But since churn is an unavoidable situation, then the question turns into finding the customer's buying patterns, content, and how that affects churn. The content of purchase and the pattern of purchase represent the relationship between consumers and supermarkets. What is the positioning of supermarkets in consumers' minds? How long will the type of relationship survive? Everyone in the retail industry will face a difficult situation. If this customer does not come, will he still survive? Or has it been lost? Because churn is not easy to define, this paper uses the context of analysis to define churn, and puts these survival data into a survival analysis model to estimate the purchase pattern and the relationship between content and survival time. In this paper, a supermarket with a Japanese chain brand is used as the research object.

六. 參考文獻
[2] 李永山(Yung-Sun Lee), 謝逸凡(Yi-Fan Hsieh) “網際網路服務業客戶流失預測模式之 研究”, Electronic Commerce Studies, vol.5, no.4, pp.485-502, 31 Dec 2007.
[1] 苦行者,Vulture, 漫草寒煙,村姑,A. Roh, funwmy, “MBA 智庫百科”,
https://wiki.mbalib.com/zh-
tw/%E5%AE%A2%E6%88%B7%E5%85%B3%E7%B3%BB%E7%AE%A1%E7%90%86.
[3] V. L. Miguéis, A. Camanho, and J. F. e Cunha, "Customer attrition in retailing: An
application of Multivariate Adaptive Regression Splines ", Qual. Reliab. Engng. Int.2004;
20:397– 406, vol. 40, no. 16, 15 Nov 2013, pp. 397-406.
[4] W.Buckin, D. V. Poel, ” Customer base analysis: partial defection of behaviourally loyal
clients in a non-contractual FMCG retail setting”, European journal of operational
research, vol.164, no.1, pp. 252—268,2005.
[5] V. L. Miguéis, A. Camanho, and J. F. e Cunha, "Customer attrition in retailing: An
application of Multivariate Adaptive Regression Splines ", Expert Systems with
Applications, vol. 40, no. 16, 15 Nov 2013, pp. 6225-6232.
[6] A. Dingli, V.Marmara, and N. S. Forunier, "Comparison of Deep Learning Algorithms to
Predict Customer Churn within a Local Retail Industry", International Journal of Machine
Learning and Computing , vol. 7, no. 5, Oct. 2017.
[7] B, Meglicki, "Linear time DBSCAN for sorted 1D data and laser range scan segmentation",
arXiv preprint arXiv:2103.16245, 30 Mar 2021.
[8] K. Khan, S. U. Rehman, K. Aziz, S. Fong and S. Sarasvady, "DBSCAN: Past, present and
future," The Fifth International Conference on the Applications of Digital Information and
Web Technologies (ICADIWT 2014), 2014, pp. 232-238, doi: 10.1109/ICADIWT.2014.6814687.
28
[9] S. J. POCOCK, S. M. GORE, G. R. KERR,” Long term survival analysis: the curability of breast cancer”, Statistics in medicine, vol.1, no.2, pp. 93—104, 1982.
[10] J. Lu, O. Park,” Modeling Customer Lifetime Value Using Survival Analysis − An Application in the Telecommunications Industry”, Data Mining Techniques, pp. 120—128, 2003.
[11] Temporal MDS Plots for Analysis of Multivariate Data - Scientific Figure on ResearchGate.
Available from: https://www.researchgate.net/figure/One-dimensional-DBSCAN- algorithm-for-similarity-values-using-a-user-defined-threshold-t_fig2_282528978 [accessed 15 Jul, 2022]