2023. 9. 15. 00:07ㆍGoogle ML Bootcamp/3. Structuring Machine Learning Projects
if Train / Dev set distribution are similar, variance problem
if Train / Dev set distribution are not similar (means come from another dataset), can't decide to variance problem.
In left case, Train-dev error 9%
- variance problem exist. cauze Train-dev set distribution is similar with Train set.
In right case, Train-dev error 1.5%, Dev error 10%
- data mismatch problem. not variance problem.
In left case, avoidable bias problem exist. not variance and data mismatch problem.
In right case, avoidable bias and data mismatch problem exist. not variance problem.
if dev error 와 Test error가 차이가 많이나는 경우, dev에 과적합되었다.
- 해결법은 Dev set을 늘리는 것. 데이터를 더 수집해야한다.
**정리 : avoidable bias 는 bigger model or more data 등등, variance는 regularization, change NN model 등등**
- data mismatch 문제는 어떻게 해결해야 하나? 다음영상에서 알아보자.
'Google ML Bootcamp > 3. Structuring Machine Learning Projects' 카테고리의 다른 글
18. Transfer Learning (0) | 2023.09.15 |
---|---|
17. Addressing Data Mismatch (0) | 2023.09.15 |
15. Training and Testing on Different Distributions (0) | 2023.09.14 |
14. Bulid your First System Quickly, then Iterate (0) | 2023.09.14 |
13. Cleaning Up Incorrectly Labeled Data (0) | 2023.09.14 |