计算机系统应用
計算機繫統應用
계산궤계통응용
APPLICATIONS OF THE COMPUTER SYSTEMS
2015年
5期
140-146
,共7页
云模型%相似性度量%约束函数%信任变化函数%协同过滤推荐
雲模型%相似性度量%約束函數%信任變化函數%協同過濾推薦
운모형%상사성도량%약속함수%신임변화함수%협동과려추천
cloud model%similarity measure%constraint function%trust change function%collaborative filtering recommendation
传统的协同过滤推荐算法面临严峻的数据稀疏性和推荐实时性困境，推荐质量明显不高。为提高推荐效果，首先对基于云模型的用户评分项和相似性度量方法展开研究。然后定义基于云模型的推荐系统信任约束，并改进主观信任云模型的约束函数、信任变化云模型的信任变化函数。最后提出一种基于云模型的协同过滤推荐算法。实验结果表明，相比传统算法，该算法在用户评分数据稀疏的状况下仍然可以取得良好的推荐效果，具有较高的实用价值。
傳統的協同過濾推薦算法麵臨嚴峻的數據稀疏性和推薦實時性睏境，推薦質量明顯不高。為提高推薦效果，首先對基于雲模型的用戶評分項和相似性度量方法展開研究。然後定義基于雲模型的推薦繫統信任約束，併改進主觀信任雲模型的約束函數、信任變化雲模型的信任變化函數。最後提齣一種基于雲模型的協同過濾推薦算法。實驗結果錶明，相比傳統算法，該算法在用戶評分數據稀疏的狀況下仍然可以取得良好的推薦效果，具有較高的實用價值。
전통적협동과려추천산법면림엄준적수거희소성화추천실시성곤경，추천질량명현불고。위제고추천효과，수선대기우운모형적용호평분항화상사성도량방법전개연구。연후정의기우운모형적추천계통신임약속，병개진주관신임운모형적약속함수、신임변화운모형적신임변화함수。최후제출일충기우운모형적협동과려추천산법。실험결과표명，상비전통산법，해산법재용호평분수거희소적상황하잉연가이취득량호적추천효과，구유교고적실용개치。
The traditional collaborative filtering recommendation algorithms face the dilemma of severe data sparsity and real time of recommendation, their recommendation quality is not obviously high. To improve recommendation efficiency, firstly, user rating items and similarity measurement method based on cloud model were researched. Then the definition of recommendation system trust constraint based on cloud model was given, and improved the constraint function of subjective trust cloud model and trust change function of trust change cloud model. Finally, a collaborative filtering recommendation algorithm based on cloud model was put forword. The experimental results show that the algorithm still obtains good recommendation efficiency on situation of user rating data sparsity compared to the traditional algorithms, it has high utility.
