Learn Before
What happens when a model has both high bias and high variance?
Question: For a supervised learning model that has both high bias and high variance, how does its performance compare on the training data and the validation data?
Sample answer: It fits the training data badly and does even worse on the validation data.
Key points:
- Training performance is poor.
- Validation performance is poorer still.
Rubric: The answer should say that the model performs poorly on the training data and performs even worse on the validation data.
0
1
Tags
Machine Learning
Deep Learning
Machine Learning Strategy
Supervised Learning
Dive into Deep Learning @ D2L
Data Science
Machine Learning Yearning @ DeepLearning.AI
Related
A classifier shows about 15% bias and 15% variance. Which description fits best?
True or False: When a classifier shows both high bias and high variance, the usual overfitting/underfitting labels do not cleanly describe the situation.
A classifier with high bias usually performs _____ on the training data.
Identify the Bias and Variance Pattern
Put the diagnostic steps in order for deciding whether a model has high bias, high variance, or both.
In the worked bias–variance diagnosis, what were the two estimated error rates?
True or False: A model that has both high bias and high variance can be overfitting and underfitting at the same time.
Equal Bias and Variance in a Fraud Detector
Match each model-performance pattern to the most likely diagnosis.
Reasoning About Bias and Variance From Training and Validation Results
Explain why a model can show both underfitting and overfitting at the same time.
Diagnosing a Fraud Model with 12% Training Error and 27% Dev Error
What happens when a model has both high bias and high variance?