Hi Mehmet! Thanks for the comment! The question about the dataset depends on its size and amount of variables. Using quantum simulators packages in Python I was able to successfully test QML models with 10K rows and nearly 10 variables. More than that starts to get costly to run the model, but not impossible. And it also depends on how you build your quantum circuit and the amount of qubits you are using. Also, the amount of parameters in your model is important to weight.

Regarding results, I was able to obtain slightly better results with quantum models in some cases, but also worse results. For the types of problem I worked with, I still wasn't able to see a huge advantage. But I would like to be able to test a case where the quantum advantage is very clear.