• get more training examples
  • try smaller sets of feature
  • try getting additional features
  • try adding polynomial features
  • try increasing/decreasing λ\lambda

Evaluating a hypothesis

将数据集划分成训练集和测试集,大约按7:3进行划分。划分的时候注意要随机划分。

Model Selection and training/validation/test sets

Advice for Applying Machine Learning
将假设的次幂d也看做是一个参数。然后模型选取的过程如下:

  1. 每个假设(d不相同)通过训练集来学习θ\theta
  2. 对每个假设得到交叉验证集的误差J(θ)cvJ(\theta)_{cv},选择最小的J(θopt)J(\theta_{opt})。这一步实际是对交叉验证集的拟合,所以为了避免过拟合,需要进行第三部
  3. 在测试集上得到测试集误差J(θopt)J(\theta_{opt}),即当前模型的评估。

训练集:交叉验证集:测试集 = 6:2 : 2

diagnosing bias vs. variance

Advice for Applying Machine Learning

Regularization and bias/variance

Advice for Applying Machine Learning
Advice for Applying Machine Learning
Advice for Applying Machine Learning

Learning Curves

Advice for Applying Machine Learning
Advice for Applying Machine Learning
Advice for Applying Machine Learning

相关文章:

  • 2021-07-09
  • 2022-01-14
  • 2022-01-21
  • 2021-08-10
  • 2021-09-08
  • 2021-10-02
  • 2022-12-23
  • 2021-08-22
猜你喜欢
  • 2021-08-14
  • 2021-05-31
  • 2021-11-04
  • 2021-06-04
  • 2022-12-23
  • 2021-04-23
相关资源
相似解决方案