Machine Learning
JinnyR
https://github.com/JinnyR
展开
-
哪些机器学习算法需要进行特征缩放 - feature scaling
通常以距离或者相似度(例如标量积scaler product)作为计算量的算法: 例如KNN, SVM。而基于概率图模型(graphical model)的算法:Fisher LDA ,Naive Bayes, Decision trees 和 Tree-based 集成方法 (RF, XGB)不会受到特征缩放的影响。Reference: https://stats.stackexchang...原创 2019-07-25 16:10:33 · 738 阅读 · 0 评论 -
z score vs. min-max scaling 优缺点
Min-max:所有特征具有相同尺度 (scale) 但不到处理outlierMin-max normalization: Guarantees all features will have the exact same scale but does not handle outliers well.Z-score normalization: Handles outliers, but do...原创 2019-07-31 16:58:15 · 5307 阅读 · 0 评论