Project 2 at Metis Data science, Codenamed Luther, uses web scraping, linear regression, and supervised machine learning techniques to look at an interesting question related to movie and film.
The Roger-Ebertron (v0.3) is a machine learning algorithm designed to learn how renowned late movie critic Roger Ebert would review movies today.
With the abundance of data on movies, scripts, reviews, and forums for everyone - critics and general audience to express their opinion on movies - can I try to model Mr. Ebert's movie ratings against everyone else's?
So I would like to explore the question: Can we model the movie preferences of the late Mr. Roger Ebert and estimate how he would continue to rate movies today?
Furthermore, can we create a machine learning algorithm that can not only rate movies similarly, but maybe even analyze movie elements and write reviews as he had?
- presentation.html - Presentation in HTML format. Click here to view in browser.
- 3 iPython Notebook files documenting the step by step process for this project
- Project write up is also available on my blog