Size of test dataset: 80
Percentage of successful automatic results: 60%
Percentage of less successful automatic results: 40%