Need a tutorial about a Regression problem in R

I really need a tutorial about a Regression problem in R with the following points:

  • there are numeric and categorical variables
  • predicting a numeric variable
  • show how to do the normalization of the data
  • training a neural network
  • analyze the results with the most popular metrics (MAPE, MSE, MAE, etc)
  • show how to predict the target variable given a new observation

Note: My main concern here is that the new observation needs to be normalized in the same way as the training data so the predicted value makes sense.

I know there are many tutorials about similar things in R but I didn't find any yet that combines the 6 points above.

I really need this. I'm like in an emergency.

Thank you in advance!

Hi @davidesp,

that's quite a few things to cover indeed. I am not sure you will find something that is tailored to exactly this list of questions. For the topic of modeling, data normalization, I can recommend that you have a look at recipes and rsample (and more generally into tidymodels). When it comes to neural networks, that is a big topic in itself and depends on your level - one of the most powerful tools out there is the keras library which also has an R interface (https://keras.rstudio.com).

For all points except number four (neural nets), please refer to the chapter on regression in "Introduction to statistical learning with R" by Witten, James, Hastie and Tibshirani. The book is available for free at the following link:
http://faculty.marshall.usc.edu/gareth-james/

For shallow neural nets, find a tutorial using R package neuralnet. For deep neural networks, see: