Pros and Cons
Pros
- Very effective with highly dimensional data.
- Works extremely well when there is a clear margin of separation.
- Effective when there are more dimensions than number of samples.
- Outliers have less of an impact as the hyperplane is affected only by the support vectors.
Cons
- Selecting an appropriate kernel can be computationally expensive/need to know the dataset very well to be able to pick the right kernel.
- Can take a large amount of time with a large dataset.
- Struggles with performance when there is a lot of overlap between the target classes or noise in classification problems.