Supervised text scoring

Question

I have a short text (subject line of an email) and I have a value which indicates the success it (people opening that email).

What is the best process and algorithm to feed the data into a machine learning experiment to generate a predictive experiment to score the future text?

Answer 1

First of all you should convert your text info the feature space and than you cant use any suitable machine learning algorithm to build classifier.

IMHO the best algorithm to calculate features from text is Word2Vec https://en.wikipedia.org/wiki/Word2vec from Google.

Answer 2

Azure ML has built-in text modules for featurization. The ones you should look at are Proprocess Text and N-gram Featuerizer.