nuget
This method uses Gradient descent algorithm. It uses number of small steps (iterations) And with each step use New theta values which results in smaller cost function value. After a while it comes to local minima (minimum cost function).
Logistic
01 Syntax
Microsoft.VisualBasic.Data.Bootstrapping.Logistic.Logistic
02 Methods
| Name | Overloads | Summary |
|---|---|---|
| sigmoid | 1 | Sigmoid function. Formula: g = 1 ./ (1 + (exp(-1 .* z))); |
03 Properties
| Name | Overloads | Summary |
|---|---|---|
| ALPHA | 1 | the learning rate |
| ITERATIONS | 1 | the number of iterations |
04 Fields
| Name | Overloads | Summary |
|---|---|---|
| theta | 1 | the weight to learn |
05 Members
sigmoid(
NumericMatrix)Sigmoid function. Formula: g = 1 ./ (1 + (exp(-1 .* z)));
Parameters
| Name | Type | Description |
|---|---|---|
z | NumericMatrix | Matrix, for which elements sigmoid function is calculated. |
Returns
Matrix with elements from sigmoid function.
ALPHA
the learning rate
ITERATIONS
the number of iterations
theta
the weight to learn