机器学习 5 machine learning advice

最新推荐文章于 2024-01-24 21:40:00 发布

原创最新推荐文章于 2024-01-24 21:40:00 发布 · 771 阅读

1 ·

CC 4.0 BY-SA版权

文章标签：

#optimization #features #training #lambda #email

Data Mining 专栏收录该内容

47 篇文章

订阅专栏

本文深入探讨了在机器学习领域中如何通过使用Bayesian正则化、在线学习和随机梯度下降等方法来优化算法性能。详细介绍了通过增加训练样本、尝试不同特征集大小、添加邮件特征以及调整迭代次数和优化算法策略来解决过拟合和欠拟合问题的方法。同时，文章还提到了转换为支持向量机作为一种调整优化目标的方法。

1. Bayesian regularization

2. online learning

stochastic gradient descent: 随机梯度下降

3. ML advice

a. more training examples => fix high variance
b. Trying a smaller set of features => fixes high variance.
c. Trying a larger set of features => fix high bias.
d. adding email features => fix high bias.
e. run gradient descent for more iterations => fixes the optimization algorithm

f. try Newton's method => fixes the optimization algorithm
g. using a different value for lambda => fixes the optimization objective
h. changing to an SVM is also another way of trying to fix the optimization objective

see more in http://download.youkuaiyun.com/detail/nomad2/3759561