- Linear models are very powerful for multidimensional data.
- In other words, it is important to be cautious of overfitting (Validation Methods for Machine Learning).
- A good summary can be found in [Getting Started with Machine Learning in Python] on page 67.
- Adjusting the regularization parameter, alpha or C, is important.
- Linear models are generally very fast, both in training and prediction.
- The prediction method is easy to understand.
- One challenge is that it is difficult to understand the meaning of coefficients.
- Linear models perform well when the number of features is greater than the number of samples.
- Here is a link explaining the difference between logistic regression and SVM.
- SVM tends to create a margin around the separating line, while logistic regression does not.
#supervised Learning #Machine Learning #Getting Started with Machine Learning in Python