nuget
This class contains routines which perform parameter selection for a model which uses C-SVC and an RBF kernel.
ParameterSelection
01 Syntax
Microsoft.VisualBasic.MachineLearning.SVM.ParameterSelection
02 Methods
03 Fields
| Name | Overloads | Summary |
|---|---|---|
| NFOLD | 2 | Default number of times to divide the data. |
| MIN_C | 2 | Default minimum power of 2 for the C value (-5) |
| MAX_C | 2 | Default maximum power of 2 for the C value (15) |
| C_STEP | 2 | Default power iteration step for the C value (2) |
| MIN_G | 2 | Default minimum power of 2 for the Gamma value (-15) |
| MAX_G | 2 | Default maximum power of 2 for the Gamma Value (3) |
| G_STEP | 2 | Default power iteration step for the Gamma value (2) |
| Threads | 2 | Used to control the degree of parallelism used in grid exploration. |
04 Members
GetList(
Double, Double, Double)Returns a logarithmic list of values from minimum power of 2 to the maximum power of 2 using the provided iteration size.
Parameters
| Name | Type | Description |
|---|---|---|
minPower | Double | The minimum power of 2 |
maxPower | Double | The maximum power of 2 |
iteration | Double | The iteration size to use in powers |
Returns
A list of the values 2 ^ p for each power p in the requested range.
Performs a Grid parameter selection, trying all possible combinations of the two lists and returning the combination which performed best. The default ranges of C and Gamma values are used. Use this method if there is no validation data available, and it will divide it 5 times to allow 5-fold validation (training on 4/5 and validating on 1/5, 5 times).
Parameters
| Name | Type | Description |
|---|---|---|
problem | Problem | The training data |
createParams | Func(Of Parameter) | The parameters to use when optimizing |
report | Action(Of GridSquare) | Function used to report results |
C | Double | The optimal C value will be put into this variable |
Gamma | Double | The optimal Gamma value will be put into this variable |
Returns
A list of grid squares and their results
Grid(Problem,
Func(Of Parameter), List(Of Double), List(Of Double), Action(Of GridSquare), Double, Double)Performs a Grid parameter selection, trying all possible combinations of the two lists and returning the combination which performed best. Use this method if there is no validation data available, and it will divide it 5 times to allow 5-fold validation (training on 4/5 and validating on 1/5, 5 times).
Parameters
| Name | Type | Description |
|---|---|---|
problem | Problem | The training data |
createParams | Func(Of Parameter) | The parameters to use when optimizing |
CValues | List(Of Double) | The set of C values to use |
GammaValues | List(Of Double) | The set of Gamma values to use |
report | Action(Of GridSquare) | Function used to report results |
C | Double | The optimal C value will be put into this variable |
Gamma | Double | The optimal Gamma value will be put into this variable |
Returns
A list of grid squares and their results
Grid(Problem,
Func(Of Parameter), List(Of Double), List(Of Double), Action(Of GridSquare), Int32, Double, Double)Performs a Grid parameter selection, trying all possible combinations of the two lists and returning the combination which performed best. Use this method if validation data isn't available, as it will divide the training data and train on a portion of it and test on the rest.
Parameters
| Name | Type | Description |
|---|---|---|
problem | Problem | The training data |
createParams | Func(Of Parameter) | The parameters to use when optimizing |
CValues | List(Of Double) | The set of C values to use |
GammaValues | List(Of Double) | The set of Gamma values to use |
report | Action(Of GridSquare) | Function used to report results |
nrfold | Int32 | The number of times the data should be divided for validation |
C | Double | The optimal C value will be placed in this variable |
Gamma | Double | The optimal Gamma value will be placed in this variable |
Returns
A list of grid squares and their results
Performs a Grid parameter selection, trying all possible combinations of the two lists and returning the combination which performed best. Uses the default values of C and Gamma.
Parameters
| Name | Type | Description |
|---|---|---|
problem | Problem | The training data |
validation | Problem | The validation data |
createParams | Func(Of Parameter) | The parameters to use when optimizing |
report | Action(Of GridSquare) | Function used to report results |
C | Double | The optimal C value will be placed in this variable |
Gamma | Double | The optimal Gamma value will be placed in this variable |
Returns
A list of grid squares and their results
Grid(Problem, Problem,
Func(Of Parameter), List(Of Double), List(Of Double), Action(Of GridSquare), Double, Double)Performs a Grid parameter selection, trying all possible combinations of the two lists and returning the combination which performed best.
Parameters
| Name | Type | Description |
|---|---|---|
problem | Problem | The training data |
validation | Problem | The validation data |
createParams | Func(Of Parameter) | The parameters to use when optimizing |
CValues | List(Of Double) | The C values to use |
GammaValues | List(Of Double) | The Gamma values to use |
report | Action(Of GridSquare) | Function used to report results |
C | Double | The optimal C value will be placed in this variable |
Gamma | Double | The optimal Gamma value will be placed in this variable |
Returns
A list of grid squares and their results
NFOLD
Default number of times to divide the data.
MIN_C
Default minimum power of 2 for the C value (-5)
MAX_C
Default maximum power of 2 for the C value (15)
C_STEP
Default power iteration step for the C value (2)
MIN_G
Default minimum power of 2 for the Gamma value (-15)
MAX_G
Default maximum power of 2 for the Gamma Value (3)
G_STEP
Default power iteration step for the Gamma value (2)
Threads
Used to control the degree of parallelism used in grid exploration. Default value is the number of processors.
Threads
NFOLD
MIN_C
MAX_C
C_STEP
MIN_G
MAX_G
G_STEP