I want to evaluate the sequential tone detection system. Although the results are similar to what is expected, the problem is that the data size is different between the predicted data and the actual data, which makes the evaluation using confusion matrix or f1-score should not be possible. So, what evaluation method is recommended for this scenario?
