andreaspb / aa-final-project-stuff Goto Github PK

We need to export all our plots into report/img/

Key points to mention:

Converting data to parametric by using vectorization - count and tf-idf
K-means
Random forest
SVM
Naive Bayes
Supervised learning

Check methods section as to not repeat something already explained

since our data is really unbalanced, we should use Youden Index to get a better idea of how we are doing

For some reason, we have empty review texts which we are cleaning up after the fact in a lot of the other scripts... This should just happen in the cleanup script

We can use TF-IDF which looks at term frequency, which we can use for our learning algorithm

Classification report
Youden's Index
Other neat stuff :^)