How can I train a neural network to grade the user's answers to a questionnaire?

Question

I have a questionnaire consisting of over 10 questions. The questionnaire is being answered by a lot of people, which I have manually graded. Each question can give the user up to 10 points depending on how they have answered.

Let's say that my dataset is big enough, how would I go about using a neural network to automatically grade these questions for me?

I have used a convolutional neural network before for image classification, but, when dealing with text classification, where should I start? Is there some sort of tutorial out there that covers this with a similar example?

Answer 1

use the embedding layers which are mainly used for text.

input the question number and the text that the student wrote to the algorithm. Make the problem into a regression of a number from 0 to 10 ( or a classification of class 10, see which one gives better performance). So the NN will look at text given a question number and try to figure out which 'points' should it get. It will be a supervised training problem since you already have examples of text rated from 0 to 10.