新型工业化
新型工業化
신형공업화
New Industrialization Straregy
2012年
11期
12-17
,共6页
计算机应用%推荐系统%奇异值分解%Slope One
計算機應用%推薦繫統%奇異值分解%Slope One
계산궤응용%추천계통%기이치분해%Slope One
computer application%recommend system%Single Value Decomposition%Slope One
Slope One 算法是一种基于项目的协同过滤推荐算法，利用线性回归模型来预测评分。Slope One算法在数据稀疏的情况下推荐效果不好，而矩阵分解在抗稀疏性方面有很有效，因此本文引入奇异值分解方法改进Slope One算法。实验结果表明，该方法有效地解决了Slope One在数据的稀疏情况下推荐效果较差的问题，提高了推荐质量。
Slope One 算法是一種基于項目的協同過濾推薦算法，利用線性迴歸模型來預測評分。Slope One算法在數據稀疏的情況下推薦效果不好，而矩陣分解在抗稀疏性方麵有很有效，因此本文引入奇異值分解方法改進Slope One算法。實驗結果錶明，該方法有效地解決瞭Slope One在數據的稀疏情況下推薦效果較差的問題，提高瞭推薦質量。
Slope One 산법시일충기우항목적협동과려추천산법，이용선성회귀모형래예측평분。Slope One산법재수거희소적정황하추천효과불호，이구진분해재항희소성방면유흔유효，인차본문인입기이치분해방법개진Slope One산법。실험결과표명，해방법유효지해결료Slope One재수거적희소정황하추천효과교차적문제，제고료추천질량。
Slope One algorithm is item-based collaborative filtering algorithm. It uses linear regression model to predict value. Slope One algorithm recommendation result is worse than the other CF algorithm under the condition of sparse data. However, matrix decomposition have great effect in sparseness resistance. In this paper, single value decomposition is used to improve Slope One algorithm. The experimental results show that Slope One algorithm based on single value decomposition sloves the problem that the effect of Slope One recommendation is worse, and improve the quality of recommendation.
