美文网首页
机器学习:5.1 模型组合 Model Combination

机器学习:5.1 模型组合 Model Combination

作者: Cache_wood | 来源:发表于2022-04-15 15:39 被阅读0次

@[toc]

Bias & Variance Decomposition

  • Learn \hat{f}_D from dataset D sampled from y= f(x) +\varepsilon

  • Evaluate generalization error (y-\hat{f}_D(x))^2 on a new data point (x,y)
    E_D[(y-\hat{f}_D(x))^2] = E_D[((f-E_D[\hat{f}_D])-(\hat{f}_D -E_D[\hat{f}_D])+\varepsilon)^2]\\ = (f-E_D[\hat{f}_D])^2+E_D[(\hat{f}_D-E_D[\hat{f}_D])^2]+\varepsilon^2\\ = Bias[\hat{f}_D]^2+Var[\hat{f}_D]+\varepsilon^2

Reduce Bias & Variance

  • Reduce bias
    • A more complex model
      • e.g. increase layers, hidden units of MLP
      • Boosting, Stacking
  • Reduce variance
    • A simpler model
      • e.g. regularization
      • Bagging, Stacking
  • Reduce \sigma^2
    • Improve data
  • Ensemble learning: train and combine multiple models to improve predictive performance

相关文章

网友评论

      本文标题:机器学习:5.1 模型组合 Model Combination

      本文链接:https://www.haomeiwen.com/subject/ebeosrtx.html