推荐方法-1:UserCF&ItemCF
Summary of recommended methods(1) 1.Metrics RMSE MAE Coverage Diversity Recall Precision 1.1 RMSE(均根方误差) R M S E = ∑ u , i ∈ T ( r u i − r ^ u i ) 2 ∣ T ∣ RMSE = \frac{\sqrt{\sum_{u, i \in T}(r_{ui} - \hat{r}_{ui}})^{2}}{\lvert T \rvert} R M S E = ∣ T ∣ ∑ u , i ∈ T ( r u i − r ^ u i ) 2 u u u : 用户 u u u i i i : 物品 i i i r u i r_{ui} r u i : 用户 u u u 对 i i i 的评分 r ^ u i \hat{r}_{ui} r ^ u i : 算法预测的评分 1.2 MAE(平均绝对值误差) M A E = ∑ u , i ∈ T ∣ r u i − r ^ u i ∣ ∣ T ∣ MAE = \frac{\sum_{u,i \in T} \lvert r_{ui} - \hat{r}_{ui} \rvert}{\lvert T \rvert} M A E = ∣ T ∣ ∑ u , i ∈ T ∣ r u i