📄 中文摘要
对KNN风格的回归模型进行了基准测试，比较了SmartKNN与多种经典KNN变体在多个真实世界数据集上的表现。测试使用了14个回归数据集，所有模型均采用默认设置，没有进行数据集特定的调优。最终排名基于平均R²得分。比较的模型包括SmartKNN、曼哈顿KNN、KDTree KNN、BallTree KNN、距离KNN、均匀KNN和切比雪夫KNN。结果显示SmartKNN在性能上优于其他经典KNN变体。
SmartKNN与经典KNN的回归基准结果
出处: SmartKNN vs Classical KNN: Regression Benchmark Results
发布: 2026年3月26日
📄 English Summary
SmartKNN vs Classical KNN: Regression Benchmark Results
A benchmark test was conducted on KNN-style regression models, comparing SmartKNN against various classical KNN variants across multiple real-world datasets. The test utilized 14 regression datasets, with all models running under default settings and no dataset-specific tuning applied. The final ranking was based on the average R² score. The models compared included SmartKNN, KNN (Manhattan), KNN (KDTree), KNN (BallTree), KNN (Distance), KNN (Uniform), and KNN (Chebyshev). The results indicate that SmartKNN outperforms other classical KNN variants in terms of performance.