Week 11: Traveling

May 05, 2019

I spent half of week 11 traveling with my family, so progress was a bit slower. Once I got back home, I started fine-tuning the randomForest network. Unfortunately, I wasn’t able to raise my accuracy past 64% for randomForest with the covtype dataset. Therefore, after consulting my external advisors, I decided to use ranger, which is a much faster black box neural network, in place of it. After putting together the initial code, I left the fine-tuning to Week 12.

Since I was having trouble with the adaboost black box model, I dug into the fastadaboost package to see what the issue was. It turns out that the fastadaboost package was primarily for binary classification (classifying something out of only two options), whereas I needed something that would work for multiclass classification. I found that adabag was a package that also ran adaboost and worked for multiclass classification. However, since my R was out of date, I had to update that first. After playing around with adaboost (from the adabag package) a bit, I determined that this wasn’t a great fit for my project either since the accuracies were pretty low.

I hope to get ranger up to a good accuracy in Week 12!

See you then,

Arshia

One Reply to “Week 11: Traveling”

  1. Vidur Gupta says:

    Looks great. Also, I was wondering what was the covtype dataset?

Leave a Reply

Your email address will not be published.