吳玲玲Wu, Ling-Ling臺灣大學:資訊管理學研究所蔣宗恩Chiang, Tzung-EnTzung-EnChiang2010-05-052018-06-292010-05-052018-06-292009U0001-3007200910552100http://ntur.lib.ntu.edu.tw//handle/246246/180034近年來網際網路上的商品數量與種類快速的增加，如何幫助消費者找到符合他們需求的商品變成網路賣家一個日漸重要的議題，而推薦系統似乎是一個可行的解決方式。而推薦系統對市場銷售造成的影響有兩派不同的說法，有學者認為推薦系統會造成熱門產品更加熱門，冷門產品趨於冷門。而另一派學者認為推薦系統會幫助消費者找到符合他們個人需求的商品，但這些商品並不一定是熱門商品，導致熱門商品的部分銷售會轉移至冷門商品上。我們發現市場上銷售集中程度的高低變化取決我們所採用的推薦策略，當推薦系統以推薦熱門商品為主，銷售集中程度上升，使的熱門商品更加熱賣。相反的，當推薦系統能夠依個人的選擇行為進行推薦，銷售集中程度下降，部分熱門商品的銷售轉移至冷門商品。而使用者個人的認知行為與對推薦系統的信任程度，並不會影響銷售集中程度改變的方向，但會影響變動的大小。因此，推薦系統有可能造成銷售集中程度的上升或下降，在考慮市場上銷售集中度的變化時，必須同時考量推薦策略、使用者的認知行為與對推薦系統的信任程度，才能夠準確的預測。In recent years, it has seen an extraordinary increase in the number of products available on the Internet. Thus, it has become increasingly important to help consumers locate desirable products from Internet. Recommenders are useful tools to solve this problem. However, there are two different views about recommenders. Some researchers believe that with the help of recommendation systems, the concentration of sales on a small number of hits will decrease. On the contrary, contradicting views that believe recommendation systems make popular products become more popular and vice-versa for unpopular ones exist. Our results indicate change of sales concentration is depended on which recommendation strategies we adopt. Sales concentration will increase when recommenders incline to recommend popular products. On the contrary, sales concentration will decrease when recommenders promote products which fit consumers’ awareness behaviors. We also add consumer’s awareness behaviors and acceptance rate into discuss. We find these two factors only change the magnitude of the effects of recommenders to sales concentration but not the direction. According to these results, we can combine awareness behaviors, recommendation strategies and acceptance rate when we predict change of sales concentration.Content 1able 2igure 3. INTRODUCTION 4. PRIOR WORK 8.1 Winner-Take-All and Long Tail Theory 8.2 Initial Product Selection 11.3 Recommendation Strategy 14.4 Decision Making 16. PROBLEM DEFINITION 18.1 Measure of Sales Concentration 18. MODEL DESIGN 20.1 Model Definition 20.2 Initial Product Selection 20.3 Recommendation Strategy 23.4 Decision Making 27. SIMULATIONS 29.1 Data Translation 29.1.1 Analysis User preference 29.1.2 Preference Function 32.1.3 Popularity Function 33.2 Initial Product Selection 34.3 Recommendation Strategy 38.4 Decision Making 39. RESULTS 40.1 Sample Path of No recommendation 41.2 Simulation Results 42. CONCLUSIONS 53.1 Contributions 55.2 Limitations 56.3 Future Research 57. REFERECE 58PPENDIX A 61application/pdf771435 bytesapplication/pdfen-US商業經濟推薦系統協同式推薦模擬長尾理論贏家通吃銷售集中Businesseconomicselectronic commercerecommender systemscollaborative filteringwinner-take-allsimulationlong tailconcentrationhttp://ntur.lib.ntu.edu.tw/bitstream/246246/180034/1/ntu-98-R96725028-1.pdf