Error rate refers to the proportion of mistakes when we apply our model to the data. Training error rate is the error rate on the training dataset. Testing error rate refers to the error rate for testing data.