I had to perform a binary classification, and from the beginning I started thinking about using the Random Forest classifier. But now I'm thinking, if using a neural network would've not been better.

So my question is, which model you would've chosen.? and if not the Neural Networks, what are the tasks the most suited for the Neural Network and why.?


There is no best model for binary classification. Instead it is better to focus on what kind of data you have and let that guide your choice of model.

Decision tree based models usually perform very well when dealing with regular tabular data of categorical and numerical features. Mainly random forest and gradient boosting (LightGBM and XGBoost). If you have very few samples a SVM can also be a good option.

Neural networks shine on more unstructured data like images and text. Time series are a bit more nuanced and often multiple models can work well.

| improve this answer | |
  • 1
    $\begingroup$ That definitely answered my question , thank you a lot . $\endgroup$ – Dimi May 24 '19 at 16:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.