PrologMethods/Classification/linear_svm.md

+# Linear SVM a L2-regularized support vector machine.
+
+An implementation of linear SVM for multiclass classification.
+
+# Available Predicates
+
+* [initModelWithTrain/10](/PrologMethods/Classification/linear_svm#initmodelwithtrain10)
+* [initModelNoTrain/4](/PrologMethods/Classification/linear_svm#initmodelnotrain4)
+* [classify/8](/PrologMethods/Classification/linear_svm#classify8)
+* [classifyPoint/3](/PrologMethods/Classification/linear_svm#classifypoint3)
+* [computeAccuracy/6](/PrologMethods/Classification/linear_svm#computeaccuracy6)
+* [train/8](/PrologMethods/Classification/linear_svm#train8)
+
+---
+
+[links/resources](/PrologMethods/Classification/linear_svm#connected-linksresources)
+
+## **_initModelWithTrain/10_**
+
+Initializes the linear_svm model with the given data and trains it.
+
+```prolog
+%% part of the predicate definition
+initModelWithTrain(  +pointer(float_array), +integer, +integer, 
+                     +pointer(float_array), +integer, 
+                     +integer, +float32, +float32, 
+                     +integer, 
+                     +string).
+```
+
+### Parameters
+| Name | Type | Description | Default |
+|------|------|-------------|---------|
+| data | +matrix | Input training features. Each column associate with one sample | - |
+| labels | +vector | Labels associated with the feature data. | - |
+| numClasses | +integer | Number of classes for classification. | 2 |
+| lambda | +float | L2-regularization constant. | 0.0001 |
+| delta | +float | Margin of difference between correct class and other classes. | 1.0 |
+| fitIntercept | +integer(bool) | add intercept term or not. | (0)false |
+| optimizer | +string | Desired optimizer: "lbfgs", "psgd" | lbfgs |
+
+---
+
+## **_initModelNoTrain/4_**
+
+Initializes the linear_svm model with the given data but doesnt train it.
+
+```prolog
+%% part of the predicate definition
+initModelNoTrain(  +integer, +float32, +float32, 
+                   +integer).
+```
+
+### Parameters
+| Name | Type | Description | Default |
+|------|------|-------------|---------|
+| numClasses | +integer | Number of classes for classification. | 2 |
+| lambda | +float | L2-regularization constant. | 0.0001 |
+| delta | +float | Margin of difference between correct class and other classes. | 1.0 |
+| fitIntercept | +integer(bool) | add intercept term or not. | (0)false |
+
+---
+
+## **_classify/8_**
+
+Classify the given points, returning class scores and predicted class label for each point.
+
+```prolog
+%% part of the predicate definition
+classify(  +pointer(float_array), +integer, +integer, 
+           -pointer(float_array), -integer, 
+           -pointer(float_array), -integer, -integer).
+```
+
+### Parameters
+| Name | Type | Description | Default |
+|------|------|-------------|---------|
+| data | +matrix | Matrix of data points to be classified. | - |
+| labels | -vector | Predicted labels for each point. | - |
+| scores | -matrix | Class probabilities for each point. | - |
+
+---
+
+## **_classifyPoint/3_**
+
+Classify the given point.
+
+```prolog
+%% part of the predicate definition
+classifyPoint(  +pointer(float_array), +integer, 
+                [-integer]).
+```
+
+### Parameters
+| Name | Type | Description | Default |
+|------|------|-------------|---------|
+| point | +vector | Point to be classified. |  |
+| - |  |  |  |
+| predictedLabel | -integer | Predicted class label of the point. | - |
+
+---
+
+## **_computeAccuracy/6_**
+
+Computes accuracy of the learned model given the feature data and the labels associated with each data point.
+
+Predictions are made using the provided data and are compared with the actual labels.
+
+```prolog
+%% part of the predicate definition
+computeAccuracy(  +pointer(float_array), +integer, +integer, 
+                  +pointer(float_array), +integer, 
+                  [-float32]).
+```
+
+### Parameters
+| Name | Type | Description | Default |
+|------|------|-------------|---------|
+| testData | +matrix | Matrix of data points using which predictions are made. | - |
+| testLabels | +vector | Vector of labels associated with the data. | - |
+| accuracy | -float | Accuracy of the model. | - |
+
+---
+
+## **_train/8_**
+
+Train the Linear_svm model with the given training data.
+
+```prolog
+%% part of the predicate definition
+train(  +pointer(float_array), +integer, +integer, 
+        +pointer(float_array), +integer, 
+        +integer, 
+        +string, 
+        [-float32]).
+```
+
+### Parameters
+| Name | Type | Description | Default |
+|------|------|-------------|---------|
+| data | +matrix | Input training features. Each column associate with one sample | - |
+| labels | +vector | Labels associated with the feature data. | - |
+| numClasses | +integer | Number of classes for classification. | 2 |
+| optimizer | +string | Desired optimizer: "lbfgs", "psgd" | lbfgs |
+| finalPointValue | -float | Objective value of the final point | - |
+
+---
+
+# Connected Links/Resources
+
+If you want a more detailed explanation, then go to the python documentation. There is most of the time a good explanation on how the methods work and what the parameters do.
+
+* MLpack::linear_svm_C++\_documentation
+* [MLpack::linear_svm_Python_documentation](https://www.mlpack.org/doc/stable/python_documentation.html#linear_svm)
+
+added some of the links from the python documentation
+
+* [random_forest](https://gitlab.cs.uni-duesseldorf.de/stups/abschlussarbeiten/prolog-mlpack-libary/-/wikis/PrologMethods/Classification/random_forest)
+* logistic_regression
+* [LinearSVM on Wikipedia](https://en.wikipedia.org/wiki/Support-vector_machine)
\ No newline at end of file