CONCEPT
- Statistical method to predict real valued data
- Relates one dependent variable with one independent variable
- Works on multiple powers of same independent variable
- Expresses the relationship using a curved line
- Non-linear model, but linear regression: depends on a linear combination of terms
KEY FEATURES
- Models non-linearly separable data
- More flexible,can work on complex relationships
- Full control over feature variables
- Requires knowledge of data and careful design
- Chances of overfitting
CODE
DATASET - Click to get the dataset
CODE - Click to get the code
No comments:
Post a Comment