I want to train a classifier to distinguish between three visual tasks based on recorded eye movements specific features.
According to what I have read so far, the training set consists of instances with their appropriate labels.
My question is, how do I determine the appropriate label, is it done manually during experiment, when I ask the user to perform a certain visual task and I record the measurements and label the instances knowningly?
Yes, for training you need the labels of the samples you give to the classifier. This labels must be correct, but how you determine them depends on your experimental setup.
Yes, manually in your experimental setup.
Then you cut de data in two parts (train data and test data). The train data is used for classifier induction. And the test data for classifier evaluation.
- Badges
- Science topic