Machine Learning Model

[BigBadButterfree]

So I'm in a machine learning class in college and for my final project I'm making a model that will predict the popularity of a given villager. As a basis I'm using the large data sheet that's been passed around with the villagers' subtypes and favorite colors and things like that. I then manually put in every villager's rank, and the model is currently only predicting at about 30% accuracy. I want to add some more fields (unfortunately probably also manually) that might help. One I have in mind is the animal's color, like marina being pink or Wendy being predominantly blue. What other fields do you think might help determine popularity? Also, for any other CS nerds out there, I'm using Weka but I'm a total noob and kinda suck at this class so if anyone has suggestions on classifier methods I should be using to improve the model please let me know

 

[mayortiffany]

One field could be whether the villager is new (to account for Raymond, Judy, Audie, and Sherb). Another could be if the villager is an Amiibo, if you're counting those. And maybe another to account for 'mythical' or 'special design' villagers like Ankha, Lucky, or Coco.

 

[BigBadButterfree]

Hm... That's a good idea. I think I'll add one for "debut game" and I don't really know how I'd determine the "special design" cuz I'd have to draw some line there.