Please, I want to clarify something. I have 5 different dataset project named as Chart, Closure, Lang, Math, and Time. The largest of these dataset is Closure. Can I use the closure to train a machine learning algorithm and use the other data set to validate the performance. Does this make any sense or is it acceptable? Thanks.
@Tao Tao,
Thanks so much sir. I am just curious if it is acceptable in fault localization as I have never seen such in the literature. Scholars do use K-Folds, but I don't want to the use K-Folds.
@Tao Tao,
Thanks so much for your response sir. I really appreciate it.
But I thought training a model with Dataset A and testing the model in dataset B,C,D, etc is Tweed as transfer of learning.
I aim to use SVM anyway. Is there a spectacular transfer of learning algorithm? If yes, could you suggest one for me?
As long as both the datasets are sampled from same distribution, you can do it. Given that condition is satisfied, one may also leverage the power of transfer learning, as already discussed in the conversation.
I think Akshay Raina has expressed it well. And those two datasets must reflect the same business logic even dimension in matter of variables
I think it would be all right to do that, if test classes as same as train classes.
Yes, if you can assume the train dataset and the test dataset belong to the same dataset.
- Badges
- Science topic
- Similar topics
- Organizational Psychology
- Training