# Data simulation using make_classification in Python

I have a question about data simulation in Python. I deal with the classification of imbalanced data and want to test the effectiveness of different methods on simulated data. I have seen in various articles and books that the make_classification function is used to generate data. Then the data is generated from a normal distribution, so the data is continuous and not discrete. Are such data correct for classification (SVM, Decision Trees) research?

There is no obstacle to doing this. For example you can create data by make_classification, and compare different algorithms by building model on it. You can also pass a random_state value to obtain same data each time you call the function. Both SVM, and Decision Trees can work with continuous data.

• Yes. I understand that. I just wonder if comparing different models built on continuous data will be good. Because in reality, often the data is categorical. May 3 '21 at 14:15
• It is quite superficial of course. Using more realistic data would be more appropriate to judge algorithms. Nevertheless you can see how well different algorithms do on continuous data. I generally use it to learn properly implement some methods, or pipeline. Not for comparing methods, but learning how to code them. May 3 '21 at 14:47
• Maybe you know any more real ways to simulate data? May 3 '21 at 14:52
• It depends on the domain of problem. For example in robotics, simulations are used to generate data to create reinforcement learning policies. Agent in the simulation generates data in each step to learn its policy. May 3 '21 at 15:00
• You can use public datasets to compare, and experiment with different algorithms. It can close the gap between superficial data, and problems in real data. May 3 '21 at 15:02