知识卡片

普通读书笔记卡 · 6484.c · 生成模型:GPT-6

案例排名需要可比协议

理解等级:高级

本章表2给出若干算法的评分误差与F1,但没有完整展开可复现的数据切分、调参和F1阈值。表中SVD误差最低,KNNBasic的F1略高,已说明“最好”取决于指标。图10却重复了误差数值,不能据图判断F1;这些材料不支持把案例名次外推为算法通用优劣。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      methods for the desired recommendation system. The RMSE score of these methods varied from 7 to 8. For improvement, the use of co-clustering, SVD, NMF, KNNbasic, KNNwithMeans, and KNNwithZScore models is done. The use of the above algorithms allowed a dramatic improvement of RMSE and MAE errors. The following table shows the RMSE value, the MAE value, and the F1 score for the implemented algorithms.

    参考来源

    - 位置:《Artificial Intelligence and Data Science in Recommendation System: Current Trends, Technologies, and Applications》第01章《推荐系统的机器学习方法》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Artificial Intelligence and Data Science in Recommendation System: Current Trends, Technologies, and Applications/_epub-src/OEBPS/Text/9789815136746-C1.html) - 结论依据:摘录中关于案例排名需要可比协议的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:methods for the desired recommendation system. The RMSE score of these methods varied from 7 to 8. For improvement, the use of co-clustering, SVD, NMF, KNNbasic, KNNwithMeans, and KNNwithZScore models is done. The use of the above algorithms allowed a dramatic improvement of RMSE and MAE errors. The following table shows the RMSE value, the MAE value, and the F1 score for the implemented algorithms.