I’m looking for a ML solution to what I thought was a simple problem. I have say 20K product codes with descriptions and I’m trying to create a ML Model that will allow me to input a description and the model picks the most suitable product code. I’ve attempted this in Azure Machine Learning Studio , transform data > selecting columns > edit meta data > process text > Train Model ( Multiclass Neural Network) > Score model . My question is this the right methodology ? Is there a better way about things ?

​

I’m very much a ML dummy so please forgive my ignorance.