近年來電子商務的興起，人們的消費行為也逐漸在改變，從傳統的實體店面購物轉變為更便捷的網路電子購物，這種消費行為的改變使的傳統實體零售業(Retail)在銷售宣傳的模式上必須有所改變。由於Email Flyers(在台灣稱作Electronic Direct-Mail)的部分，成本較低廉，因此很多傳統實體零售業會藉由大量的發送電子促銷DM來吸引顧客回到門市進行消費，將最近的促銷品皆放入電子DM之中。就消費者行為而言，當顧客回到門市進行消費時，可能會購買不在預定購買清單或是電子DM中的商品，所以發送電子DM的主要目的是藉由DM吸引顧客回到門市進行消費。但電子DM中卻不一定都是顧客所喜好的商品，將過多種類的商品放入於電子DM以及發送次數過於頻繁時，將導致顧客必須使用相當多的時間來閱讀電子DM才能找到自己所喜好的商品；可能會導致顧客對於電子DM的觀感變差，產生厭惡感，因此無法吸引顧客回到門市進行消費。在本研究中，將利用協同過濾推薦系統的分析，設計一個適合以食品為主超市的推薦演算法，依據Cross-selling的概念，針對『顧客已經購買過的商品』和『顧客未購買過的商品』兩個因素進行考量，將顧客最有可能購買的商品製作成客製化電子DM，藉此吸引顧客回到門市進行消費。因此，本研究中所提出的推薦演算法除了能顧及顧客原有的喜好，藉此吸引顧客回到門市進行消費外，還能將更多顧客購買先前未購買過的商品推薦給顧客，以增加超市收益。

With the rise of e-commerce in recent years, most people have changed their purchase behavior. Instead of going to the physical stores shopping, people prefer to buy things online conveniently. These changes of purchase behavior causing traditional retail trade must change the way their advertising pattern. Due to the cheaper cost of email flyer (or electronic Direct-Mail), traditional retail trade would send lots of e-DM to attract customers to return back to physical retail store and put promotion merchandise into e-DM to attract customers to return to physical stores. When customers return to stores and buy merchandises, they probably would buy some merchandise that weren’t on the shopping list or on the e-DMs, therefore, the main purpose sending e-DM is attracting customers back to stores and purchase merchandises. However, not every merchandises on e-DM was customers’ favorite merchandise, putting too many kinds of merchandise or sending too many e-DMs would make customers spend too much time on finding their favorite merchandise on e-DMs. This might be leaving customers a bad impression, and stopped them from returning to the stores for shopping. In this paper, we would design a proper algorithm by analyzing Collaborative filtering recommender system for supermarket. According to the concept of cross-selling, we would consider these two factors, the merchandise that customers had bought and the merchandise that they hadn’t bought, and choose the most possible merchandise to make customize e-DM to attract customers return to store and purchase merchandise. Therefore, the algorithm we designed in this article could not only considering the customers’ purchase behavior to attract them return to store and purchase merchandise but also recommending the merchandise that customers hadn’t bought before to increase revenue for supermarket.

目錄
第1章 緒論 1
1.1 研究背景與動機	1
1.2 研究架構 4
第2章 相關文獻與研究探討 5
2.1 以內容為基礎的推薦系統 6
2.2 協同過濾推薦系統 8
2.3 混合式推薦系統	17
第3章 研究流程 19
3.1 問題陳述 19
3.2 研究方法 26
第4章 實驗結果與探討 31
4.1 對所有顧客進行推薦 33
4.1.1 使用COS、BIG和CBF演算法 33
4.1.2 推薦新商品類別與重複購買相同商品類別分析 36
4.2 對目標顧客進行推薦 44
4.2.1 使用COS、BIG和CBF演算法 44
4.2.2 推薦新商品類別與重複購買相同商品類別分析 47
4.3 使用CBF組合演算法 52
4.3.1 對目標顧客進行推薦 52
4.3.2 對所有顧客進行推薦 62
第5章 結論與未來研究方向 68
參考文獻 71
附錄-英文論文 76
 
圖目錄
圖1 顧客對於商品的評價分數矩陣 10
圖2 Cos User-based計算範例 13
圖3 Cos Item-based計算範例 14
圖4 顧客商品評分表 20
圖5 Cosine Correlation Coefficient商品之間的相關性 20
圖6 顧客對於未購買過商品的相關性 21
圖7 預測顧客對於未購買過的商品可能會產生的評價 21
圖8 顧客交易紀錄 25
圖9 二元矩陣商品類別之間的相關性 25
圖10 Content-based Filtering Algorithm 30
圖11 所有顧客推薦4個商品類別的推薦成功率圖 34
圖12 所有顧客推薦5個商品類別的推薦成功率圖 34
圖13 所有顧客推薦6個商品類別的推薦成功率圖 35
圖14 所有顧客推薦4個商品類別的新商品類別推薦成功率圖 37
圖15 所有顧客推薦5個商品類別的新商品類別推薦成功率圖 37
圖16 所有顧客推薦6個商品類別的新商品類別推薦成功率圖 38
圖17 所有顧客推薦4個商品類別的重複購買相同商品類別推薦成功率圖 39
圖18 所有顧客推薦5個商品類別的重複購買相同商品類別推薦成功率圖 39
圖19 所有顧客推薦6個商品類別的重複購買相同商品類別推薦成功率圖 40
圖20 目標顧客推薦4個商品類別的推薦成功率圖 45
圖21 目標顧客推薦5個商品類別的推薦成功率圖 45
圖22 目標顧客推薦6個商品類別的推薦成功率圖 46
圖23 目標顧客推薦4個商品類別的新商品類別推薦成功率圖 47
圖24 目標顧客推薦5個商品類別的新商品類別推薦成功率圖 48
圖25 目標顧客推薦6個商品類別的新商品類別推薦成功率圖 48
圖26 目標顧客推薦4個商品類別的重複購買相同商品類別推薦成功率圖 49
圖27 目標顧客推薦5個商品類別的重複購買相同商品類別推薦成功率圖 50
圖28 目標顧客推薦6個商品類別的重複購買相同商品類別推薦成功率圖 50
圖29 目標顧客推薦4個商品類別的推薦成功率圖 54
圖30 目標顧客推薦5個商品類別的推薦成功率圖 54
圖31 目標顧客推薦6個商品類別的推薦成功率圖 55
圖32 目標顧客推薦4個商品類別的新商品類別推薦成功率圖 57
圖33 目標顧客推薦5個商品類別的新商品類別推薦成功率圖 57
圖34 目標顧客推薦6個商品類別的新商品類別推薦成功率圖 58
圖35 目標顧客推薦4個商品類別的重複購買相同商品類別推薦成功率圖 60
圖36 目標顧客推薦5個商品類別的重複購買相同商品類別推薦成功率圖 60
圖37 目標顧客推薦6個商品類別的重複購買相同商品類別推薦成功率圖 61
圖38 所有顧客推薦4個商品類別的推薦成功率圖 63
圖39 所有顧客推薦5個商品類別的推薦成功率圖 63
圖40 所有顧客推薦6個商品類別的推薦成功率圖 64
圖41 所有顧客推薦4個商品類別的新商品類別推薦成功率圖 64
圖42 所有顧客推薦5個商品類別的新商品類別推薦成功率圖 65
圖43 所有顧客推薦6個商品類別的新商品類別推薦成功率圖 65
圖44 所有顧客推薦4個商品類別的重複購買相同商品類別推薦成功率圖 66
圖45 所有顧客推薦5個商品類別的重複購買相同商品類別推薦成功率圖 66
圖46 所有顧客推薦6個商品類別的重複購買相同商品類別推薦成功率圖 67
 
表目錄
表1 以內容為基礎的推薦系統的主要問題表	7
表2 顧客交易紀錄2011年1月至8月 32
表3 商品類別銷售Top10統計圖(1與2月) 41
表4 商品類別銷售Top10統計圖(3與4月) 42
表5 商品類別銷售Top10統計圖(5與6月) 42
表6 商品類別銷售Top10統計圖(7與8月) 43

[1]	張淑俐, "超市賣場推薦系統實作," PhD Unpublished Thesis, 淡江大學, 2013.
[2]	紀政宏, "推薦系統的架構與實作," PhD Unpublished Thesis, 淡江大學, 2011.
[3]	M. Pazzani and D. Billsus, "Content-based recommendation systems," The adaptive web, pp. 325-341, 2007.
[4]	N. J. Belkin and W. B. Croft, "Information filtering and information retrieval: two sides of the same coin?," Commun. ACM, vol. 35, pp. 29-38, 1992.
[5]	R. Baeza-Yates and B. Ribeiro-Neto, "Modern information retrieval vol. 463: ACM press New York.," 1999.
[6]	G. Salton, Automatic text processing: the transformation, analysis, and retrieval of information by computer: Addison-Wesley Longman Publishing Co., Inc., 1989.
[7]	M. F. Porter, "An algorithm for suffix stripping," Program: electronic library and information systems, vol. 14, pp. 130-137, 1993.
[8]	J. J. Rocchio, "Relevance feedback in information retrieval," 1971.
[9]	P. W. Foltz and S. T. Dumais, "Personalized information delivery: An analysis of information filtering methods," Communications of the ACM, vol. 35, pp. 50-60, 1992.
[10]	S. Loeb, "Architecting personalized delivery of multimedia information," Communications of the ACM, vol. 35, pp. 39-47, 1992.
[11]	M. d. Gemmis, P. Lops, G. Semeraro, and P. Basile, "Integrating tags in a semantic content-based recommender," presented at the Proceedings of the 2008 ACM conference on Recommender systems, Lausanne, Switzerland, 2008.
[12]	V. Zanardi and L. Capra, "Social ranking: uncovering relevant content using tag-based recommender systems," presented at the Proceedings of the 2008 ACM conference on Recommender systems, Lausanne, Switzerland, 2008.
[13]	CiteULike. Available: http://www.citeulike.org/home
[14]	X. N. Lam, T. Vu, T. D. Le, and A. D. Duong, "Addressing cold-start problem in recommendation systems," presented at the Proceedings of the 2nd international conference on Ubiquitous information management and communication, Suwon, Korea, 2008.
[15]	A. I. Schein, A. Popescul, L. H. Ungar, and D. M. Pennock, "Methods and metrics for cold-start recommendations," presented at the Proceedings of the 25th annual international ACM SIGIR conference on Research and development in information retrieval, Tampere, Finland, 2002.
[16]	C. Wartena, W. Slakhorst, and M. Wibbels, "Selecting keywords for content based recommendation," presented at the Proceedings of the 19th ACM international conference on Information and knowledge management, Toronto, ON, Canada, 2010.
[17]	G. Adomavicius and A. Tuzhilin, "Toward the Next Generation of Recommender Systems: A Survey of the State-of-the-Art and Possible Extensions," IEEE Trans. on Knowl. and Data Eng., vol. 17, pp. 734-749, 2005.
[18]	U. Shardanand and P. Maes, "Social information filtering: algorithms for automating “word of mouth&rdquo," presented at the Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, Denver, Colorado, USA, 1995.
[19]	D. Billsus and M. J. Pazzani, "User Modeling for Adaptive News Access," User Modeling and User-Adapted Interaction, vol. 10, pp. 147-180, 2000.
[20]	E. Rich, "User modeling via stereotypes," in Readings in intelligent user interfaces, T. M. Mark and W. Wolfgang, Eds., ed: Morgan Kaufmann Publishers Inc., 1979, pp. 329-342.
[21]	D. Goldberg, D. Nichols, B. M. Oki, and D. Terry, "Using collaborative filtering to weave an information tapestry," Commun. ACM, vol. 35, pp. 61-70, 1992.
[22]	U. Shardanand and P. Maes, "Social information filtering: algorithms for automating "word of mouth"," presented at the Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, Denver, Colorado, USA, 1995.
[23]	J. A. Konstan, B. N. Miller, D. Maltz, J. L. Herlocker, L. R. Gordon, and J. Riedl, "GroupLens: applying collaborative filtering to Usenet news," Commun. ACM, vol. 40, pp. 77-87, 1997.
[24]	L. Terveen and W. Hill, "Beyond Recommender Systems: Helping People Help Each Other," ed, 2001.
[25]	G. Linden, B. Smith, and J. York, "Amazon.com recommendations: item-to-item collaborative filtering," Internet Computing, IEEE, vol. 7, pp. 76-80, 2003.
[26]	R. M. Bell and Y. Koren, "Lessons from the Netflix prize challenge," SIGKDD Explor. Newsl., vol. 9, pp. 75-79, 2007.
[27]	B. Pradel, S. Sean, J. Delporte, S. Guerif, C. Rouveirol, F. Fogelman-Soulie, et al., "A case study in a recommender system based on purchase data," presented at the Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining, San Diego, California, USA, 2011.
[28]	J. Wang, B. Sarwar, and N. Sundaresan, "Utilizing related products for post-purchase recommendation in e-commerce," presented at the Proceedings of the fifth ACM conference on Recommender systems, Chicago, Illinois, USA, 2011.
[29]	J. B. Schafer, J. Konstan, and J. Riedl, "Recommender systems in e-commerce," presented at the Proceedings of the 1st ACM conference on Electronic commerce, Denver, Colorado, USA, 1999.
[30]	M. Deshpande and G. Karypis, "Item-based top-N recommendation algorithms," ACM Trans. Inf. Syst., vol. 22, pp. 143-177, 2004.
[31]	B. Sarwar, G. Karypis, J. Konstan, and J. Riedl, "Item-based collaborative filtering recommendation algorithms," presented at the Proceedings of the 10th international conference on World Wide Web, Hong Kong, Hong Kong, 2001.
[32]	L. Candillier, F. Meyer, and M. Boull, "Comparing State-of-the-Art Collaborative Filtering Systems," presented at the Proceedings of the 5th international conference on Machine Learning and Data Mining in Pattern Recognition, Leipzig, Germany, 2007.
[33]	X. Su and T. M. Khoshgoftaar, "A survey of collaborative filtering techniques," Adv. in Artif. Intell., vol. 2009, pp. 2-2, 2009.
[34]	C. G. Gonzalez, J. W. Bonventi, and A. L. Rodrigues, "Density of Closed Balls in Real-Valued and Autometrized Boolean Spaces for Clustering Applications," presented at the Proceedings of the 19th Brazilian Symposium on Artificial Intelligence: Advances in Artificial Intelligence, Savador, Brazil, 2008.
[35]	D. Lemire and A. Maclachlan, "Slope One Predictors for Online Rating-Based Collaborative Filtering," in Proceedings of the 2005 SIAM International Conference on Data Mining, ed, 2005, pp. 471-475.
[36]	H. C. Kum, J. H. CHANG, and W. WANG, "Sequential pattern mining in multi-databases via multiple alignment," Data Mining and Knowledge Discovery, vol. 12, pp. 151-180, 2006.
[37]	D. Billsus and M. J. Pazzani, "Learning Collaborative Information Filters," presented at the Proceedings of the Fifteenth International Conference on Machine Learning, 1998.
[38]	T. K. Landauer and M. L. Littman, "Computerized cross-language document retrieval using latent semantic indexing," ed: Google Patents, 1994.
[39]	S. C. Deerwester, S. T. Dumais, T. K. Landauer, G. W. Furnas, and R. A. Harshman, "Indexing by latent semantic analysis," JASIS, vol. 41, pp. 391-407, 1990.
[40]	P. Resnick and H. R. Varian, "Recommender systems," Communications of the ACM, vol. 40, pp. 56-58, 1997.
[41]	K. Pearson, "LIII. On lines and planes of closest fit to systems of points in space," The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science, vol. 2, pp. 559-572, 1901.
[42]	M. Zanker, "A collaborative constraint-based meta-level recommender," in Proceedings of the 2008 ACM conference on Recommender systems, 2008, pp. 139-146.
[43]	A. I. Schein, A. Popescul, L. H. Ungar, and D. M. Pennock, "Methods and metrics for cold-start recommendations," in Proceedings of the 25th annual international ACM SIGIR conference on Research and development in information retrieval, 2002, pp. 253-260.
[44]	Q. Li and B. M. Kim, "Clustering approach for hybrid recommender system," in Web Intelligence, 2003. WI 2003. Proceedings. IEEE/WIC International Conference on, 2003, pp. 33-38.
[45]	R. Burke, "Hybrid web recommender systems," in The adaptive web, ed: Springer, 2007, pp. 377-408.
[46]	A. Gunawardana and C. Meek, "A unified approach to building hybrid recommender systems," in Proceedings of the third ACM conference on Recommender systems, 2009, pp. 117-124.
[47]	G. Lekakos and P. Caravelas, "A hybrid approach for movie recommendation," Multimedia tools and applications, vol. 36, pp. 55-70, 2008.
[48]	MovieLens. Available: http://www.movielens.org
[49]	Y.-H. Lee, T.-H. Cheng, C.-W. Lan, C.-P. Wei, and P. J.-H. Hu, "Overcoming small-size training set problem in content-based recommendation: a collaboration-based training set expansion approach," in Proceedings of the 11th International Conference on Electronic Commerce, 2009, pp. 99-106.
[50]	S. H. Choi, Y.-S. Jeong, and M. K. Jeong, "A hybrid recommendation method with reduced data for large-scale application," Systems, Man, and Cybernetics, Part C: Applications and Reviews, IEEE Transactions on, vol. 40, pp. 557-566, 2010.
[51]	C. Kim and J. Kim, "A Recommendation Algorithm Using Multi-Level Association Rules," presented at the Proceedings of the 2003 IEEE/WIC International Conference on Web Intelligence, 2003.
[52]	B. Sarwar, G. Karypis, J. Konstan, and J. Riedl, "Analysis of recommendation algorithms for e-commerce," presented at the Proceedings of the 2nd ACM conference on Electronic commerce, Minneapolis, Minnesota, USA, 2000.
[53]	朱奐禎, "結合時間因素預測消費者回訪與回購機率," PhD Unpublished Thesis, 淡江大學, 2014.