摘要: 1. Fit locally If the true model changes much, we want to fit our function locally to different regions of the input space. 2. Scaled distance \ we pu 阅读全文
posted @ 2017-05-07 18:23 ClimberClimb 阅读(87) 评论(0) 推荐(0) 编辑
摘要: 1. Feature selection Sometimes, we need to decrease the number of features Efficiency: With fewer features, we can compute quickly Interpretaility: wh 阅读全文
posted @ 2017-05-07 10:44 ClimberClimb 阅读(146) 评论(0) 推荐(0) 编辑