Today I continued with the IBM intro ML course, and got to build a multilinear regression model. This time I tried to do it with data outside of the class. I found some data on IMDB movies that also had a Metacritic score and a number of other features. Below is a scatter graph of just the IMDB and metascore ratings.
We then go through the usual steps of preparing the data and running a regression but this time, I had an extra dependant variable. You can see it in the code below, line 29.
This model had a variance score of just 0.58 so it’s not of much use. But it was fun to test something I’d just learnt and with a different data set. Below, I run the model with Dune’s IMDB rating from 2021.
The model predicted a Metacritic score of 71. The actual rating today is 74. But you don’t need a fancy model to know that IMDB ratings are usually in line with Metacritic scores!