在九十年代中期，推薦系統開始被學者們關注進而研究。推薦系統的主要任務即是找出客戶的潛在需求與行為，並透過這些資訊幫助企業或組織提高產品銷售量或客戶服務品質。現今，主要有三種不同的推薦系統，分別為Contents-based、Collaborative-filtering 和 Hybrid recommendation。其中Collaborative-filtering推薦系統是最為普及的類型，但仍存在一些限制。第一，其無法依序進行推薦，例如：買床之後應推薦床單。第二，此類型的推薦系統必須在大量使用者的情況之下才能進行有效的推薦。而在此篇論文中，我們把重心放在解決第一個限制，並提出一個序列相似度的計算方式。
在最後的實驗中我們與傳統的Collaborative-filtering推薦系統進行校能的比較，在實驗中，我們有較佳的效能。

In the mid-1990s, recommender systems have been concerned by researchers. The main task of recommender system is providing information, which can matches latent interests of customers. The recommender system is aimed to suggest and provide information of the products to customers to help them find product which they need quickly. There are mainly three kinds of methods in recommender system: Contents-based, Collaborative-filtering and Hybrid recommendation.
The collaborative filtering is the most popular and successful recommender system, it still has several limitations. First, there is no way to recommend items in sequence. Second, the success of the collaborative recommender system depends on the availability of mass users. In this paper, we focus on the first limitations and attempt to remedy this limitation of collaborative-filtering recommendation by developing the novel approach to group these sequential transactions. The key idea of our approach from the following important observation: As we know intimately, the behavior for purchasing is influenced by sequence relationship among items. For instance, such as customer may buy jelly after buying toast. It is very useful for us to understand the motivation for purchasing which is hidden behind the behavior for purchasing. It means that we need to recognize what sequential purchase behaviors is user actually follows. Then, we use that the information of customers to recommend items to customers.
Besides, we offer a new measure to compute similarity between sequences. Differing from other similarity measures, we provide a distance-sensitive similarity measure. Thus, the performance of our measure is better.

[1] Gediminas Adomavicius, Alexander Tuzhilin, “Toward the Next Generation of Recommender Systems: A Survey of the Sate-of-the-Art and Possible Extensions,” IEEE Transcations on Knowledge and Data Engineering, 2005, pp: 734-749
[2] Yeong Bin Cho, Yoon Ho Cho, Soung Hie Kim, “Mining changes in customer buying behavior for collaborative recommendations,” Expert Systems with Application, 2005, pp: 359-369
[3]Seung-Joon Oh ., Jae-Yearn Kim, “A hierarchical clustering algorithm for categorical sequence data,” Information Processing Letters ,2004, pp: 135–140
[4] Berson,A., Smith, S., and Thearling, K. Building Data Mining Applications for CRM, 1999
[5] Hirji, k. “Exploring Data Mining Implementation,” Communications of the ACM, 2001, pp: 87-93.
[6] Ansari, S., Kohavi, R.,Mason, L., and Zheng, Z. “Integrating E-Commerce and Data Mining: Architecture and Challenges,” IEEE, 2001, pp:27-34.
[7] Yasuo H., Takao T., Yukichi O. “Recommending Books of Revealed and Latent Interests in E-Commerce,” IEEE, 2000, pp: 1632-1637.
[8] Yu, S. P. “Data Mining and Personalization Technologies,” IEEE, 1999,pp: 6-13
[9] Balabanovic, M., Shoham, Y. “Fab: Content-Based, Collaborative Recommendation,” Communications of the ACM, 1997,pp: 66-72
[10] Herlocker, J., Konstan, J., and Riedl, J. “Explaining Collaborative Filtering Recommendations,” In Proceedings of CSCW ’00, 2000,pp:241-247
[11] Goldberg, D., Nichols, D., Oki, B. M., and Terry, D. “Using Collaborative Filtering to weave an Information Tapestry,” Communications of the ACM, 1992, pp: 61-70.
[12] Resnick, P., Iacovou, N., Suchak, M., Bergstrom, P., and Riedl, J. “GroupLens: An Open Architecture for Collaborative Filtering of Netnews,” In Proceedings of ACM CSCW ’94 Conference on Computer-Supported Cooperative Work, 1994, pp: 175-186.
[13] Breese, J.S., Heckerman, D., and Kadie, C. “Empirical Analysis of Predictive Algorithms for Collaborative Filtering,” In Proceedings of the 14th conference on Uncertainty in Artificial Intelligence, 1998.
[14] Zeng, C., Xing, C. X., and Zhou, L. Z. “Similarity Measure and Instance Selection for Collaborative Filtering,” Communications of the ACM, 2003, pp:652-658
[15] Sarwar, B., Konstan, J., Borchers, A., Herlocker, J., Miller, B., and Riedl, J. “Using Filtering Agents to Improve Prediction Quality into the GroupLens Research Collaborative Filtering System,” In Proceedings of CSCW ’98, 1998, pp: 1-10.
[16] Good, N., Schafer, B., Konstan, J., Borchers, A., Sarwar, B., Herlocker, J., and Riedl, J. “Combing Collaborative Filtering With Personal Agents for Better Recommendations, “ In Proceedings of the AAAI ’99 conference, 1999.
[17] Sarwar, B., Karypis, G., Konstan, J., and Riedl, J. “Analysis of Recommendation Algorithms for E-Commernce,” In Proceedings of the ACM EC ’00 Conference, 2000, pp: 158-167.
[18]Sarwar, B., Karypis, G., Konstan, J., and Riedl, J. “Application of Dimensionality Reduction Recommender System – A Case Study,” In ACM WebKDD Workshop, 2000.
[19]Sarwar, B., Karypis, G., Konstan, J., and Riedl, J. “Item-based Collaborative Filtering Recommendation Algorithms, “ Communications of the ACM, 2001, pp:285-295
[20] Jonathan L. Herlocker, Joseph A. Konstan, Al Borchers, and Riedl, J. “An algorithmic Framework for Performing Collaborative Filtering,” In Proceedings of the 22nd annual international ACM SIGIR conference, pp230-237.
[21] Pawlak, Z. Grzymala-busse, J., Slowinski, R., and Ziarko, W. “Rough Set,” Communications of the ACM, 1995, pp: 89-95.
[22] Grazymala-Busse, J., Ziarko, W. “Data Mining and Rough Set Theory,” Communication of the ACM, 2000, pp: 108-109
[23] Qiankun Zhao, Sourav S. Bhowmick, “Sequential Pattern Mining: A Survey”, Technical Report, 2003