Predicting happiness from demographics and poll answers
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
amelia
LICENSE
README.md
validation.r
vectorize_and_predict_inplace.py
vectorize_validation.py

README.md

Predicting happiness from demographics and poll answers

Code for The Analytics Edge (15.071x) competition.

amelia - impute missing values, train many models, predict, bag
validation.r - split the training set for validation, train and score random forest and naive Bayes,
	plot variable importance from random forest
vectorize_and_predict_inplace.py - convert categorical to -1/0/1, train, write predictions	
vectorize_validation.py - convert data to numbers only, train, get validation score

Get 0.74568 public / 0.77761 private AUC with vectorize_and_predict_inplace.py and even better score with Amelia.

For description, see: