Support Vector Machines (SVM) - Support Vector Classifier/Regression - Large/Maximal Margin Classifier/Regression
- is a non-parametric, instance-based, discriminative-typed, supervised learning model
- used mainly for classification problems (sometimes regression)
- has capacity to learn linear models and complex non-linear models
Support Vector Classifier
- is a hyperplane that linearly separates data into 2 categories
SVM - Model Types
main idea:
| |
|
When there is no obvious linear classifier that separates data in a nice way Non-Linear SVM works by transforming data into a high dimensional space then finding the Support Vector Classifier main idea:
|
|
Hard-Margin SVM |
|
|---|---|
|
Soft-Margin SVM |
|
SVM - Model Examples
|
Hard Margin |
Soft Margin | |
|---|---|---|
|
Linear |
|
|
|
Non-Linear |
|
|
---cognitive-computing---machine-intelligence/ai---subfields/machine-learning-(ml)---pattern-recognition/ml---models/support-vector-machines-(svm)/maximal-margin-classifier.png)
---cognitive-computing---machine-intelligence/ai---subfields/machine-learning-(ml)---pattern-recognition/ml---models/support-vector-machines-(svm)/soft-margin-classifier-or-support-vector-classifier.png)
---cognitive-computing---machine-intelligence/ai---subfields/machine-learning-(ml)---pattern-recognition/ml---models/support-vector-machines-(svm)/support-vector-machines.png)
---cognitive-computing---machine-intelligence/ai---subfields/machine-learning-(ml)---pattern-recognition/ml---models/support-vector-machines-(svm)/support-vector-machines-2.png)
---cognitive-computing---machine-intelligence/ai---subfields/machine-learning-(ml)---pattern-recognition/ml---models/support-vector-machines-(svm)/support-vector-machines-soft-margin.png)
---cognitive-computing---machine-intelligence/ai---subfields/machine-learning-(ml)---pattern-recognition/ml---models/support-vector-machines-(svm)/support-vector-machines-2-soft-margin.png)
SVM - Multi-class Classification
one-vs-all method - by training 𝑘 SVM problems (similar to multinomial logistic regression)
SVM - Other
- Logistic Regression vs Linear SVM vs Non-Linear SVM
- Support Vector Machines (SVM) as Quadratic Programming