摘要：The overload of rapid development of information is caused by the popularity of the Internet and e-commerce. Personalized recommendation system came into being. The goal of personalized recommendation is recommended the target not only meets user interest and also expands their perspective. The traditional collaborative filtering technology tends to concentrate the recommended range， thus the result is always more in line with recommended user interest， but in the diversity which is to extend the user perspective on the performance is to be improved.

This paper presents a strategy considering the capacity of the products that leads users into new area to improve the diversity of recommendation. As simply considering the diversity， it needs coordinate with other accuracy tend algorithms. This algorithm analyzes the recommended behavior of accuracy algorithms and determines the expansion by the capability that can leads user to unknown area， considering the accuracy and scalability.

關(guān)鍵詞：推薦系統(tǒng)；擴(kuò)展性；多樣性；推薦網(wǎng)絡(luò)；物品關(guān)聯(lián)網(wǎng)絡(luò)