Functions for calculating Network error. The error is simply the difference between the correct output and the actual output.
Methods
(static) arcTan(expected, actual) → {number}
Parameters:
Name | Type | Description |
---|---|---|
expected |
Array.<number> | Array of output values the Network should have produced. |
actual |
Array.<number> | Array of output values the Network actually produced. |
Returns:
- Type
- number
(static) crossEntropy(expected, actual) → {number}
Parameters:
Name | Type | Description |
---|---|---|
expected |
Array.<number> | Array of output values the Network should have produced. |
actual |
Array.<number> | Array of output values the Network actually produced. |
Returns:
- Type
- number
(static) meanSquared(expected, actual) → {number}
Parameters:
Name | Type | Description |
---|---|---|
expected |
Array.<number> | Array of output values the Network should have produced. |
actual |
Array.<number> | Array of output values the Network actually produced. |
Returns:
- Type
- number
(static) rootMeanSquared(expected, actual) → {number}
Parameters:
Name | Type | Description |
---|---|---|
expected |
Array.<number> | Array of output values the Network should have produced. |
actual |
Array.<number> | Array of output values the Network actually produced. |
Returns:
- Type
- number