https://editor.p5js.org/maxchu/sketches/9O64PO-mG
The dataset I chose to work with was the 250 top movies on IMDB. It afforded a lot of inputs for the model, but the goal was to have a simple predictor of the Metascore rating based off of runtime, language, and year released. I would think this might have been some commentary on the Film industry’s bias for older, english movies with longer runtimes.
However, there were a number of problems with the code and elsewhere that I couldn’t get around. Firstly, I was coding in VSCode, and the same code that worked in the web editor would not work in VSCode. The machine learning popup displayed and the neural network went through all of its 200 epochs, but when it came down to press the results button, nothing showed up. Further, when I retooled the code in class, it didn’t run the machine learning process at all. I worked through it for a while, but I couldn’t get it to run. I think the issue might be the fact that I’m asking the model to neglect most of the classifiers and only focus on a select three of them, which in that case I would have to do heavy editing to the csv file, but I hope that’s not the case.
I will keep working.
Leave a Reply