Predicting happiness from demographics and poll answers
Python R
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
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: