How can I train a neural network with information about academic courses (mostly text as input, 1 numeric) to classify whether the course is future proof?

Hi guys,

The test database that I have is a has 6 input variables (5 in text, 1 numeric) and one (0 or 1) output variable saved in a table. These 6 input variables describe content of academic courses. The name, description, learning outcomes, level, #credits and department. In a training set of 93 courses I identified 20 courses that are FutureProof. This is a time intensive task. The rest of the database consist of ±2000 courses.

How can I train a neural network (if this is the right choice) on this training set to identify which of the other ±1900 courses can be classified as future proof and which cannot?

Thanks a lot in advance!

Answer 1

[Vishal Bhutani]

By my understanding, you want to train a neural network which is having input data of both type numerical as well as nominal variables. The link attached will refer to the same issue you are facing:

https://www.mathworks.com/matlabcentral/answers/116666-how-to-train-a-neural-network-with-an-input-data-set-that-comprises-of-numeric-values-as-well-as-nom

Hope it helps.

[Christiaan Teeuwen]

Dear Vishal,

Thanks for your answer!

If I understand the link that you posted correctly, they categorize in the following manner:

"Oil = 1; Water = 2; Slickwater = 3; Thebloodoftheinnocent = 4;

Fluid = Oil;"

Here the distiction is clear between the words that are transformed into categories. However, the input that I try to use consists of two instances with text varying from e.g. 10-100 words that are sentences and therefore not categorizable. The output variable has a 0 or 1 output.

I hope this clarifies the problem I am having. Would you know how to such a thing?

Kind regards, Christiaan