Build a superhero api with machine learning capabilities
Dataset and exercise we use as part of the hiring process for machine learning related jobs at e-conomic (Visma).
From exercise.tex;
Here at e-conomic we really love superheros. In fact we love superheros so much that we have decided to start a new department which will focus solely on inventing new superhero characters. Our team of dedicated inventors will think out new awesome characters and for each new creation the inventor will devise a character description containing information such as abilities, partnerships, team affiliations, etc.
The two legendary US publishers, Marvel Comics and DC Comics, have shown great interest in our new department, and they would like to make offers on each new character....
Due to market circumstances, when a new character is offered to a publisher, all other publishers lose interest in buying that character, because they are afraid that the company who got the offer first, may steal the idea and bring onto the market a similar superhero, before the investment can be legally secured.
That is why we need your help! We mined Wikipedia for Marvel- and DC-superhero articles, and now we hand you the dataset.
Your task is to build a RESTful API that (i) manages the articles, (ii) and offers a prediction endpoint, that uses machine learning to classify a text as being either closer to the Marvel or the DC universe.