new LDA(…classes)
An LDA object.
Parameters:
| Name | Type | Attributes | Description |
|---|---|---|---|
classes |
Array.<Array.<number>> |
<repeatable> |
Each parameter is a 2d class array. In each class array, rows are samples, columns are variables. |
Example
let classifier = new LDA(class1, class2, class3);Methods
classify(point) → {number}
Classify an unknown point. Uses a pairwise voting system in the event of multiclass classification.
Parameters:
| Name | Type | Description |
|---|---|---|
point |
Array.<number> | The data point to be classified. |
Returns:
Returns the predicted class. Class numbers range from 0 to (number_of_classes - 1).
- Type
- number
project(point) → {number}
Project the unknown data point to one dimension.
Currently only supports binary LDA.
Parameters:
| Name | Type | Description |
|---|---|---|
point |
Array.<number> | The data point to be projected. |
Returns:
value less than 0 if predicted to be in class 1, 0 if exactly in between, greater than 0 if class 2
- Type
- number