PrecisionRecallLayer.md

CPrecisionRecallLayer Class

• CPrecisionRecallLayer Class
  • Settings
    • Resetting the data after each run
  • Trainable parameters
  • Inputs
  • Outputs
    • Getting the results

This class implements a layer that calculates the number of objects classified correctly for either class in a binary classification scenario.

Using these statistics, you can easily calculate the precision and recall for the trained network.

Settings

Resetting the data after each run

void SetReset( bool value );

Specifies if the data should be reset after each network run. By default, the reset is turned on.

If you turn off this setting, the total values since the last reset will be calculated.

Trainable parameters

This layer has no trainable parameters.

Inputs

The layer has two inputs. The first input accepts a blob with the network response, of the dimensions:

• BatchLength * BatchWidth * ListSize is equal to the number of objects that were classified.
• Height, Width, Depth, and Channels are equal to 1.

The second input should contain a blob of the same dimensions with the correct class labels (1 for one class and -1 for the other).

Outputs

The single output contains a blob of the dimensions:

• Channels is equal to 4
• all other dimensions are equal to 1

The four elements of the blob contain:

1. The number of objects of the 1 class that were classified correctly.
2. The total number of the 1 class objects.
3. The number of objects of the -1 class that were classified correctly.
4. The total number of the -1 objects.

If you have set SetReset() to false, the layer will accumulate the data for all network runs since the last reset.

Getting the results

void GetLastResult( CArray<int>& results );

Writes the four statistics into an array in the same order as for the output blob.