Skip to main content
A classifier’s quality depends on the size of the training batch. For classifier training, use all available images in a batch. However, this makes it impossible to evaluate the quality of a classifier outside of its training documents. The Benchmark tool was created for classifier testing. It lets you split the documents into training and test documents using a percentage ratio. Use the Benchmark tool only for unbiased evaluations of the classifier quality. To create the classifier itself, use all available images during training.

​
Split the batch into training and test documents

1

Open the Split Batch window

On the Batch menu, open the Split Batch window.
2

Set the split ratio

Select the appropriate ratio for splitting the source batch. The default is 60%-40%.
3

Set the minimum number of documents per class

Specify the minimum number of documents of each class in the resulting new batches. By default, each batch must have at least 1 document of each class.
4

Name the batches and split

Specify names for the batches and click Split.

​
Train and test the classifier

Use the batches created with Split for training and testing.
1

Open the Batches window

On the Batch menu, select the Batches window.
2

Select Benchmark

Select Benchmark.
3

Select the training batch

In the drop-down list, specify the batch to use for training.
4

Select the test batches

Select the check boxes of the batches to use for testing the trained classifier.
5

Run the benchmark

Click Run. Training starts, and classification starts immediately after training.

​
View the test results

To analyze the classification results of the test batch, click Statistics in the Batches window. For more information, see Classifier statistics.