tl;dr Using a Naïve Bayesian classifier and a dataset of 1515 video game ratings, I am predicting which developer is most likely to make a game with specific properties (metascore, ESRB rating, genre, platform) in the future. Naïve Bayesian learning A Naïve Bayes classifier is a very simple method to predict categorial outcomes. A well-known application is text classification, especially predicting whether a text is spam. Here, the classifier tries to use the information about the occurrence of certain words in telling us whether an e-mail message is spam or not.
There is a game I play / try to make myself okay / try so hard to make the pieces all fit / smash it apart / just for the f**k of it (Nine Inch Nails: The Big Come Down) After this rather distressing opening by the Nine Inch Nails, let’s turn to a more uplifting topic: video games! There is a dataset on Kaggle with ratings for over 4000 video games.