在台灣各項網路服務之廣告及電子商務，是入口網站主要的獲利方式，因此利用消費者在網路上留下的行為足跡，預測消費者對於廣告可能的迴響程度，進而規劃其因應行銷策略，具有其重要性。然而消費者興趣通常會隨著時間變化而改變，為了掌握會員在不同時間點的興趣差異，本研究結合『時間函數』與『消費者過去行為』兩個因素，設計一應用於消費者回訪與回購機率的基礎模型；對於不同的資料集與應用（回訪與回購率的預測）時，只需修改參數即可。我們以某知名入口網站所提供的資料集作為實驗資料，由實驗證明我們的模型確實精準找尋到高回訪與高回購潛力之會員，以供入口網站的行銷人員有效益的行銷策略，進而增加獲利。

Consumer market has several characteristics in common such as revisit over the relevant time frame, a large number of customers, and a wealth of information detailing past customer purchases. Analyzing the characterizations and temporal dependencies of purchase behaviors is crucial for the enterprise to survive in a continuously changing environment. The internet advertising revenues and the commodity sales play an import role in the earning origin of e-commerce. Therefore, monitoring the members’ browse and purchase records has become emphasized for the prediction of the advertisement. Effective advertising requires predicting how a user responds to effective advertising requires targeting (presenting the ad) in ways that reflect these users’ preferences. The behavioral target leverages historical users’ behaviors in order to select the ads which are most related to the users to display. Although we would not like to provide advertisements to customers, with the same concept, we would expect to predict return visit and repurchase rates for the registered members. However, customers’ preferences change over time. In order to capture the concept drift, we propose a novel and simple time function to increase/decrease the weight of the old data to various members’ past behaviors. In this research, we will develop a basic model for predicting the customers’ return visit and repurchase rates. The basic model can appropriately modified for the different applications (prediction for the customers’ return visit or repurchase rates) Our achievement can be used to assist the marketers to target the members with high return visit (repurchase) rates and design corresponding marketing strategies.

目錄
第1章 緒論 1
1.1 研究背景與動機	1
1.2 研究架構 6
第2章 相關文獻與研究探討 7
2.1 顧客輪廓(Customer profile) 7
2.2 概念飄移(Concept drift) 10
第3章 問題陳述 15
3.1 訓練資料集 – 一個月 17
3.2 訓練資料集 – 三個月 19
第4章 研究方法 22
4.1 考慮時間因素之點擊興趣指數 23
4.2 CPIT與回訪率之關係探討 30
4.3 CPIT模型 35
第5章 實驗結果與探討 39
5.1 CPIT model之準確性 40
5.2 CPIT model與CPI比較 44
5.3 CPIT model與CPI結合時間函數之比較 50
第6章 消費者回購機率預測探討 53
6.1 考慮時間因素之回購指數 54
6.2 RIT model的準確性 62
6.3 RIT model與RI的比較 66
6.4 RIT model與RI結合時間函數之比較 72
第7章 結論與未來研究方向 75
參考文獻 77

圖目錄
圖 1 利用不同T0所呈現的時間函數曲線 14
圖 2 CPI6各群組在七月的回訪率 17
圖 3 CPI4-6各群組在七月的回訪率 19
圖 4 The Procedure of Merging CPIT 27
圖 5 CPIT6與CPI6之比較 28
圖 6 各月份回訪率對應關係 30
圖 7 當月沒有來會員，五月份到七月份各月份回訪率對應關係 31
圖 8 當月有來會員，五月份到七月份各月份回訪率對應關係 31
圖 9 The Algorithm of the CPIT Model 35
圖 10 增益曲線於五月份之比較(CPIT4) 37
圖 11 增益曲線於六月份之比較(CPIT5) 38
圖 12 增益曲線於七月份之比較(CPIT6) 38
圖 13 各月份（八月至十一月）各月份CPIT與回訪率對應關係 40
圖 14 增益曲線於八月份之比較(CPIT7) 41
圖 15 增益曲線於九月份之比較(CPIT8) 42
圖 16 增益曲線於十月份之比較(CPIT9) 42
圖 17 增益曲線於十一月份之比較(CPIT10) 43
圖 18 CPIT7與CPI7比較 46
圖 19 CPIT與CPI比較八月回訪率之累計獲益圖 47
圖 20 CPIT與CPI比較九月回訪率之累計獲益圖 48
圖 21 CPIT與CPI比較十月回訪率之累計獲益圖 48
圖 22 CPIT與CPI比較十一月回訪率之累計獲益圖 49
圖 23 CPIT與CPI_Decay八月回訪率之累計獲益圖比較 52
圖 24 RIT6與RI6之比較 59
圖 25 各月份（5月至7月）回購率對應關係 60
圖 26 各月份（八月至十一月）各月份RIT與回訪率對應關係 62
圖 27 增益曲線於8月份之比較(RIT7) 63
圖 28 增益曲線於九月份之比較(RIT8) 64
圖 29 增益曲線於十月份之比較(RIT9) 64
圖 30 增益曲線於十一月份之比較(RIT10) 65
圖 31 RIT7與RI7比較 68
圖 32 RIT與RI比較八月回購率之累計獲益圖 69
圖 33 RIT與RI比較九月回購率之累計獲益圖 70
圖 34 RIT與RI比較十月回購率之累計獲益圖 70
圖 35 RIT與RI比較十一月回購率之累計獲益圖 71
圖 36 CPIT與CPI_Decay比較八月回訪率之累計獲益圖 74

表目錄
表 1 CPI_6第9群中，不同類型之消費者的回訪率 18
表 2 CPI6為0不同類型之消費者在七月的回訪率 18
表 3 CPI6在第7群，不同類型的消費者的回訪狀況 20
表 4 五種類型之消費者對於七月的回訪率觀察 23
表 5 兩種不同類型消費者回訪率之比例 25
表 6 兩種不同行為之範例列表 44
表 7 五種類型之消費者對於七月的回購率觀察 54
表 8 兩種不同類型消費者回購率之比例 57
表 9 兩種不同購買行為之範例列表 66

Akehurst, J., Koprinska, I., Yacef, K., Pizzato, L., Kay, J., & Rej, T. (2012). Explicit and Implicit User Preferences in Online Dating. In L. Cao, J. Huang, J. Bailey, Y. Koh & J. Luo (Eds.), New Frontiers in Applied Data Mining (Vol. 7104, pp. 15-27): Springer Berlin Heidelberg.

Albadvi, A., & Shahbazi, M. (2009). A hybrid recommendation technique based on product category attributes. Expert Systems with Applications: An International Journal, 36(9), 11480-11488. doi: 10.1016/j.eswa.2009.03.046

Balabanovic, M., & Shoham, Y. (1997). Fab: content-based, collaborative recommendation. Communications of the ACM, 40(3), 72. doi: 10.1145/245108.245124

Bell, R., Koren, Y., & Volinsky, C. (2008). The BellKor 2008 Solution to the Netflix Prize.

Blanco-FernA!ndez, Y., Pazos-Arias, J. J., LA3pez-Nores, M., Gil-Solla, A., Ramos-Cabrer, M., GarcAa-Duque, J., . . . DAaz-Redondo, R. P. (2010). Incentivized provision of metadata, semantic reasoning and time-driven filtering: Making a puzzle of personalized e-commerce. Expert Systems with Applications: An International Journal, 37(1), 61-69. 

Chen, C. C., Chen, M. C., & Sun, Y. (2001). PVA: a self-adaptive personal view agent system. Paper presented at the In Proceedings of the seventh ACM SIGKDD international conference on Knowledge discovery and data mining, New York, NY, USA.

Chen, Y., Pavlov, D., & Canny, J. F. (2010). Behavioral Targeting: The Art of Scaling Up Simple Algorithms. ACM Transactions on Knowledge Discovery from Data (TKDD), 4(4), 17. 

Cho, Y. H., & Kim, J. K. (2004). Application of Web usage mining and product taxonomy to collaborative recommendations in e-commerce. Expert Systems with Applications: An International Journal, 26(2), 233-246. doi: 10.1016/S0957-4174(03)00138-6

Cho, Y. H., Kim, J. K., & Kim, S. H. (2002). A personalized recommender system based on web usage mining and decision tree induction. Expert Systems with Applications: An International Journal, 23(3), 329-342. 

Cooley, R., Mobasher, B., & Srivastava, J. (1999). Data preparation for mining world wide web browsing patterns. Knowledge and Information Systems, 1(1), 5-32. 

Cunningham, P. a., Nowlan, N., Delany, S. J., & Haahr, M. (2003). A Case-Based Approach to Spam Filtering that Can Track Concept Drift. Paper presented at the In The ICCBR'03 Workshop on Long-Lived CBR Systems. http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.14.3235

Ding, Y., & Li, X. (2005). Time weight collaborative filtering. Paper presented at the In Proceedings of the 14th ACM international conference on Information and knowledge management.

Fan, W. (2004). Systematic data selection to mine concept-drifting data streams. Paper presented at the In Proceedings of the tenth ACM SIGKDD international conference on Knowledge discovery and data mining.

Fan, W., Gordon, M. D., & Pathak, P. (2005). Effective profiling of consumer information retrieval needs: a unified framework and empirical comparison. Decision Support Systems, 40(2), 213-233. 

Jinhyung, C., Kwiseok, K., & Yongtae, P. (2007). Collaborative Filtering Using Dual Information Sources. Intelligent Systems, IEEE, 22(3), 30-38. doi: 10.1109/mis.2007.48

Kelly, D., & Teevan, J. (2003). Implicit feedback for inferring user preference: a bibliography. SIGIR Forum, 37(2), 18-28. doi: 10.1145/959258.959260

Konstan, J. A., Miller, B. N., Maltz, D., Herlocker, J. L., Gordon, L. R., & Riedl, J. (1997). GroupLens: applying collaborative filtering to Usenet news. Communications of the ACM, 40(3), 77-87. 

Koren, Y. (2010). Collaborative filtering with temporal dynamics. Commun. ACM, 53(4), 89-97. doi: 10.1145/1557019.1557072

Lee, J., Podlaseck, M., Schonberg, E., & Hoch, R. (2001). Visualization and analysis of clickstream data of online stores for understanding web merchandising. Data Mining and Knowledge Discovery, 5(1), 59-84. doi: 10.1023/A:1009843912662

Ma, S., Li, X., Ding, Y., & Orlowska, M. E. (2007). A recommender system with interest-drifting. Paper presented at the Proceedings of the 8th international conference on Web information systems engineering, Nancy, France.

Middleton, S. E., Shadbolt, N. R., & Roure, D. C. D. (2004). Ontological user profiling in recommender systems. ACM Transactions on Information Systems, 22(1), 88. 

Min, S. H., & Han, I. (2005). Detection of the customer time-variant pattern for improving recommender systems. Expert Systems with Applications: An International Journal, 28(2), 189-199. doi: 10.1016/j.eswa.2004.10.001

Mladenic, D. (1996). Personal WebWatcher: design and implementation.

Mobasher, B. (2007). Data mining for web personalization. The Adaptive Web, 90-135. 

Nunez-Valdez, E. R., Cueva Lovelle, J. M., Sanjuan Martinez, O., Garcia-Diaz, V., Ordonez de Pablos, P., & Montenegro Marin, C. E. (2012). Implicit feedback techniques on recommender systems applied to electronic books. Computers in Human Behavior, 28(4), 1186-1193. doi: http://dx.doi.org/10.1016/j.chb.2012.02.001

Park, Y. J., & Chang, K. N. (2009). Individual and group behavior-based customer profile model for personalized product recommendation. Expert Systems with Applications: An International Journal, 36(2), 1932-1939. doi: 10.1016/j.eswa.2007.12.034

Pazzani, M. J., Muramatsu, J., & Billsus, D. (1996). Syskill & Webert: Identifying interesting web sites. Paper presented at the Proceedings of the national conference on artificial intelligence.

Peppers, D., & Rogers, M. (2005). Return on customer: creating maximum value from your scarcest resource (Vol. 5): Crown Business.

Piton, T., Blanchard, J., Briand, H., & Guillet, F. (2009). Domain driven data mining to improve promotional campaign ROI and select marketing channels. Paper presented at the Proceedings of the 18th ACM conference on Information and knowledge management. 

Rodriguez, A., Chaovalitwongse, W. A., Zhe, L., Singhal, H., & Pham, H. (2010). Master Defect Record Retrieval Using Network-Based Feature Association. Systems, Man, and Cybernetics, Part C: Applications and Reviews, IEEE Transactions on, 40(3), 319-329. 

Salganicoff, M. (1997). Tolerating Concept and Sampling Shift in Lazy Learning UsingPrediction Error Context Switching. Artificial Intelligence Review - Special issue on lazy learning, 11(Journal Article), 133-155. 

Sang Hyun, C., Young-Seon, J., & Jeong, M. K. (2010). A Hybrid Recommendation Method with Reduced Data for Large-Scale Application. Systems, Man, and Cybernetics, Part C: Applications and Reviews, IEEE Transactions on, 40(5), 557-566. 

Shavlik, J., Calcari, S., Eliassi-Rad, T., & Solock, J. (1998). An instructable, adaptive interface for discovering and monitoring information on the World-Wide Web. Paper presented at the Proceedings of the 4th international conference on Intelligent user interfaces.

Tkalčič, M., Odić, A., Košir, A., & Tasič, J. (2012). Impact of implicit and explicit affective labeling on a recommender system’s performance. Advances in User Modeling, 342-354. 

Tsymbal, A. (2004). The Problem of Concept Drift: Definitions and Related Work.

Vuk, M., & Curk, T. (2006). ROC curve, lift chart and calibration plot. Metodoloski zvezki, 3(1), 89-108. 

Weng, S. S., & Liu, M. J. (2004). Feature-based recommendations for one-to-one marketing. Expert Systems with Applications: An International Journal, 26(4), 493-508. 

Xiang, L., & Yang, Q. (2009). Time-Dependent Models in Collaborative Filtering Based Recommender System. Paper presented at the Proceedings of the 2009 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology - Volume 01. http://portal.acm.org/citation.cfm?id=1632382

Zhao, J., Yu, X., & Sun, J. (2008). TDCF: Time Distribution Collaborative Filtering Algorithm. Paper presented at the In Proceedings of the 2008 International Symposium on Information Science and Engieering.