Hey,
I don't exactly understand how I make the decision to choose a SVM or a DNN or Linear Classifier for my binary classification problem. Are there any aspects to proove and then decide? Or is it just trying out and taking the best one? I should have a guideline or anything to describe why I use a specific classifier for my problem.
I have 5 Features which lead to a binary classification.
Anybody can help me with that? Thanks
In any problem, usually before going for statistical or stochastic solution, traditional methods, such as simple searching, classifications should be tried. You have to try from its atomic stage; else, the problem would be more critical. Even before, you have to analyse whether the problemreally is NP-Complete.
It totally depends on Your Data size ...!
If your dataset contains more than 10k with 5 features, I think you can simply use End-to-End Deep Learning and try to improve only the hyperparameters ...!
Instead, if you'd a strong hand skills in ML and small dataset, I suggest to use XgBoost or SVM (SVR, SVC)
wow thank you all for your answers, they really helped me a lot! I really appreciate that!
- Badges
- Science topic