What is Matlab doing when training a K-NN classifier?
6 visualizaciones (últimos 30 días)
Mostrar comentarios más antiguos
Moritz Hesse
el 20 de Mzo. de 2019
Respondida: Rohan Amarapurkar
el 28 de Mzo. de 2019
I am currently using the Classification Learner GUI to make a K-NN classifier model. Training (training data size: 800,000x52), is seemingly taking forever. I am wondering, what is Matlab doing while training? Since K-NN uses lazy learning, I thought there is no training required? So, why is it taking so long to "train"?
If this "training phase" is actually used to test unseen data to evalutate the model accuracy, is there any way to get more verbose output to see progress instead of just a indeterminate progress bar?
0 comentarios
Respuesta aceptada
Rohan Amarapurkar
el 28 de Mzo. de 2019
The training time might depend on the type of validation that you are using. If you are using k-fold validation, and the k is large it is likely that the model will take longer to train. The training phase uses the validation data (taken from the training dataset) to evaluate the accuracy of the model.
0 comentarios
Más respuestas (0)
Ver también
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!