Practical Use of NN

Hi everyone,

I am new to DL, and would like to know if neural network is a helpful tool to train a classification dataset with only 3-5 features. If not, why? Does anyone have such experience?

Hello @Yufei_Jing

Welcome to the community.

If you have a dataset that requires a simple decision boundary to classify the data correctly then a logistic regression model would be sufficient for the task. As the decision boundary gets more and more complicated, we would need to look at more advanced models such as Neural networks.

Hi, @Yufei_Jing!

I would suggest to make a good exploratory data analysis and try to find the couple of rules that best describe the “behaviour” of your data. With so few features you can even try a simple decision tree before a neural network