本文旨在介紹，一間零售商如何在線上通路逐漸於消費者的購物習慣中流行、市場份額逐漸擴大的時下，於激烈的競爭中勝出；或是零售商為了與本身已成熟的線下通路互補發展成新零售模式時，需要找出線上通路建置的指導方針時，提供指引。
在線上通路的成功必須要要讓消費者停留的時間增長，反向思考也可以是減少消費者浪費在商品頁面搜尋的時間，更甚者是結合兩者，當零售商的線上通路推薦系統能越精準的行銷消費者想買但卻未知的商品；消費者想買卻還沒買的商品時，最基本的就是能夠增加消費者的購物品項差異提升利潤，更可以因為個性化行銷而增加消費者額外的購物時間來瀏覽零售商的線上頁面，還可以因為正向的刺激增加消費者對通路與品牌的第一印象，並得以避免線上行銷的可能產生消費者反彈的風險，進而加強忠誠度與減少排斥感。
因此，本文透過取得Instacart的線上顧客交易紀錄，來對商品採取基本的分群，一為日常生活所需的必需品、提供消費者額外效用非必需品、特殊用途或特殊場合的特殊品，二為搜尋品、經驗品、信任品來建置推薦系統並試用不同的演算法來嘗試，這些商品分群對業者來說較為簡易，能在實際運用上將業者本身的產品進行初步分類，並採用本文所得到的結果，套用適當的推薦系統演算法提高精準性，達到與競爭對手之差異化，在推播商品時有較高的成功率與較低的顧客反感。

The purpose of the essay is to introduce how a retailer can win in the fierce competition when virtual channels are gradually popular among consumers' shopping habits and the market share of virtual channels becomes greater than before. This research also provides the retailer who needs to find guidelines for its virtual channel construction in order to complement their mature offline channels into a new selling model.

With the view of the success of the virtual channel, managers must increase the amount of time that consumers spend on the website as well as reverse thinking which can reduce consumers' search cost on product pages. Managers can also combine both of the methods to accurately recommend the products that consumers don’t know they want to buy. Regarding the commodities that consumers want to buy, but they have not bought yet, the basic way for a company to profits is from differentiating the products. As for increasing the time that consumers spend on virtual channels, it is proved that personalized market can make it possible. On the other hand, it is able to promote consumers’ first impressions of the channel and the brand with positive stimulus while avoiding the risk of a consumer rebound, and thereby enhancing loyalty and reducing the sense of rejection.

Subsequently, this research obtains consumers' online transaction records through Instacart and groups the commodities that consumers had bought. The first group includes necessities, non-essential commodities that provides consumers additional utilities and products with specific purposes or used in special occasions. The second are the search products, experienced products and trusted products. In this part, we will construct a recommendation system and try different algorithms to testify them. This grouping method of commodities is a relatively simple way for businesses, and also an actual use for them to start an initial classification of their own products. Adapting the results of the essay, the managers will be able to achieve differentiation with competitors, and to increase a higher success rate and decrease the probability of customers' dislike when promoting products with appropriate recommendation system algorithms to increase the accuracy.

目錄

第一章 緒論 10
第一節 研究背景與動機 10
第二節 研究目的 12
第三節 研究流程 14
第二章 文獻探討 15
第一節 推薦系統 15
1-1協同過濾推薦系統（Collaborative filtering recommender systems） 16
1-1-1記憶與模型基礎推薦系統（Memory & Model-based recommender systems） 17
1-1-2使用者基礎協同過濾推薦系統（User-based collaborative filtering recommender systems） 18
1-1-3項目基礎協同過濾推薦系統（Item-based collaborative filtering recommender systems） 20
1-2內容基礎分析推薦系統（Content-Based recommender systems） 23
1-3知識基礎分析推薦系統（Knowledge-based recommender systems） 24
1-4混合式推薦系統（Hybrid recommender systems） 24
1-5關聯規則推薦系統（Association Rules recommender systems） 25
第二節 線上行銷（Online Marketing） 27
2-1顧客停留（Consumer Duration） 27
2-2推播反彈（Push Notification Backlash） 30
2-3針對式廣告（Targeting Advertisement） 34
第三節 產品類別（Product Category） 37
3-1 產品知識（Product knowledge） 37
3-2 必需品、非必需品與特殊品 （Necessity、Non-necessity and Speciality goods） 39
3-3 搜尋品、經驗品與信任品 （Search、Experience and Credence goods） 41
第三章 研究設計 45
第一節 研究方法 推薦演算法（Recommender Algorithms） 45
1-1推薦演算法（Recommender Algorithms）之意義與用途 45
1-2推薦演算法（Recommender Algorithms）之種類 46
1-3推薦演算法（Recommender Algorithms）於本研究之適用性 46
第二節 研究個案 47
1-1個案介紹 47
1-2個案線上通路介紹 48
第三節 研究資料 61
3-1資料介紹 61
3-2資料處理 62
3-3推薦系統模型建置 70
第四章 研究結果 73
第一節 必需品、非必需品與特殊品 （Necessity、Non-necessity and Speciality goods） 73
1-1必需品（Necessity goods） 73
1-2非必需品（Non-necessity goods） 74
1-3特殊品（Speciality goods） 75
1-4 小結 76
第二節 搜尋品、經驗品與信任品 （Search、Experience and Credence goods） 77
2-1搜尋品（Search goods） 77
2-2經驗品（Experience goods） 78
2-3信任品（Credence goods） 79
2-4 小結 80
第五章 結論與建議 82
第一節 研究結論與建議 82
第二節 研究貢獻 83
第三節 研究限制與未來發展 84
參考文獻 87
推薦系統 87
網路消費者行為 89
產品類別 90
網路資料 91

參考文獻

推薦系統
Malone, T. W., Grant, K. R., Turbak, F. A., Brobst, S. A., & Cohen, M. D. (1987). Intelligent information-sharing systems. Communications of the ACM, 30(5), 390-402.
Sarwar, B., Karypis, G., Konstan, J., & Riedl, J. (2001, April). Item-based collaborative filtering recommendation algorithms. In Proceedings of the 10th international conference on World Wide Web (pp. 285-295). ACM.
Ricci, F., Rokach, L., & Shapira, B. (2011). Introduction to recommender systems handbook. In Recommender systems handbook (pp. 1-35). springer US.
Melville, P., Mooney, R. J., & Nagarajan, R. (2002, July). Content-boosted collaborative filtering for improved recommendations. In Aaai/iaai (pp. 187-192).
Herlocker, J. L., Konstan, J. A., Borchers, A., & Riedl, J. (1999, August). An algorithmic framework for performing collaborative filtering. In Proceedings of the 22nd annual international ACM SIGIR conference on Research and development in information retrieval (pp. 230-237). ACM.
Bagley, D. (2017). Combining Content Information with an Item-Based Collaborative Filter
Gorakala, S. K., & Usuelli, M. (2015). Building a recommendation system with R. Packt Publishing Ltd.
何明, 刘伟世, & 张江. (2017). 支持推荐非空率的关联规则推荐算法. 通信学报, 38(10), 18-25.
Agrawal, R., & Srikant, R. (1994, September). Fast algorithms for mining association rules. In Proc. 20th int. conf. very large data bases, VLDB (Vol. 1215, pp. 487-499).
Özseyhan, C., Badur, B., & Darcan, O. N. (2012). An association rule-based recommendation engine for an online dating site. Communications of the IBIMA, 2012, 1.
Hahsler, M. (2011). recommenderlab: A framework for developing and testing recommendation algorithms. Southern Methodist University.
Mobasher, B., Dai, H., Luo, T., & Nakagawa, M. (2001, November). Effective personalization based on association rule discovery from web usage data. In Proceedings of the 3rd international workshop on Web information and data management (pp. 9-15). ACM.
Lin, W., Alvarez, S. A., & Ruiz, C. (2002). Efficient adaptive-support association rule mining for recommender systems. Data mining and knowledge discovery, 6(1), 83-105.
Han, J., Pei, J., & Kamber, M. (2011). Data mining: concepts and techniques. Elsevier.
Yao, G., & Cai, L. User-Based and Item-Based Collaborative Filtering Recommendation Algorithms Design.
Breese, J. S., Heckerman, D., & Kadie, C. (1998, July). Empirical analysis of predictive algorithms for collaborative filtering. In Proceedings of the Fourteenth conference on Uncertainty in artificial intelligence (pp. 43-52). Morgan Kaufmann Publishers Inc..

網路消費者行為
Buxbaum, O. (2016). Key Insights Into Basic Mechanisms of Mental Activity. Springer.
Kim, J., Fiore, A. M., & Lee, H. H. (2007). Influences of online store perception, shopping enjoyment, and shopping involvement on consumer patronage behavior towards an online retailer. Journal of retailing and Consumer Services, 14(2), 95-107.
Cho, E., & Youn-Kyung, K. (2012). The effects of website designs, self-congruity, and flow on behavioral intention. International Journal of Design, 6(2).
Malär, L., Krohmer, H., Hoyer, W. D., & Nyffenegger, B. (2011). Emotional brand attachment and brand personality: The relative importance of the actual and the ideal self. Journal of Marketing, 75(4), 35-52.
Dholakia, U. M. (2015). The Perils of Algorithm-Based Marketing. Harvard Business Review. June, 17.
Tsang, M. M., Ho, S. C., & Liang, T. P. (2004). Consumer attitudes toward mobile advertising: An empirical study. International journal of electronic commerce, 8(3), 65-78.
Campbell, D. E., & Wright, R. T. (2008). Shut-up I don't care: Understanding the role of relevance and interactivity on customer attitudes toward repetitive online advertising. Journal of Electronic Commerce Research, 9(1), 62.
Tractinsky, N., & Lowengart, O. (2007). Web-store aesthetics in e-retailing: A conceptual framework and some theoretical implications. Academy of Marketing Science Review, 2007, 1.

產品類別
林南宏, 王文正, 邱聖媛, & 鍾怡君. (2007). 產品知識及品牌形象對購買意願的影響-產品類別的干擾效果. 行銷評論, 4(4), 481-504.
Brucks, M. (1985). The effects of product class knowledge on information search behavior. Journal of consumer research, 1-16.
Rao, A. R., & Monroe, K. B. (1988). The moderating effect of prior knowledge on cue utilization in product evaluations. Journal of consumer research, 15(2), 253-264.
Babin, B. J., Darden, W. R., & Griffin, M. (1994). Work and/or fun: measuring hedonic and utilitarian shopping value. Journal of consumer research, 20(4), 644-656.
姜淳方, & 李昀修. (2012). 台灣連鎖速食餐廳屬性, 享樂及功利價值, 行為意圖關係之研究—以台灣 Y 世代消費者為例. 行銷科學學報, 8(1), 77-95.
Nelson, P. (1970). Information and consumer behavior. Journal of political economy, 78(2), 311-329.
Hao, Y., Ye, Q., Li, Y., & Cheng, Z. (2010, January). How does the valence of online consumer reviews matter in consumer decision making? Differences between search goods and experience goods. In System sciences (HICSS), 2010 43rd hawaii international conference on (pp. 1-10). IEEE.
Aggarwal, P., & Vaidyanathan, R. (2005). Perceived effectiveness of recommendation agent routines: search vs. experience goods. International Journal of Internet Marketing and Advertising, 2(1-2), 38-55.
Hsieh, Y. C., Chiu, H. C., & Chiang, M. Y. (2005). Maintaining a committed online customer: a study across search-experience-credence products. Journal of Retailing, 81(1), 75-82.
Dulleck, U., & Kerschbamer, R. (2005). On Doctors, Mechanics and Computer Specialists-The Economics of Credence Goods.
Emons, W. (1997). Credence goods and fraudulent experts. The RAND Journal of Economics, 107-119.

網路資料
eMarketer (2011). New Data on Attitudes Toward Targeting https://www.emarketer.com/Article/New-Data-on-Attitudes-Toward-Targeting/1008222
Monty Majeed (2016). How excessive marketing is killing your startup https://yourstory.com/2016/07/pitfalls-of-excessive-marketing/
Leslie K. John, Tami Kim, & Kate Barasz (2017). Ads That Don’t Overstep https://hbr.org/product/ads-that-dont-overstep/R1801C-PDF-ENG
Instacart: Grocery Delivery https://play.google.com/store/apps/details?id=com.instacart.client&rdid=com.instacart.client
Rodrigo (2012). STP: Segmentation, Targeting and Positioning in Marketing Strategies https://writepass.com/journal/2012/12/segmentation-targeting-and-positioning-in-marketing-strategies/