#ebay sale predictor
A little toy project which scrapes and prepares a dataset from ebay, trains a scikit-learn Gradient Boosting Classifier on the price and ad title, and tries to predict the probability of the listing ending in a sale. The performance of the model is actually quite good, considering that it is only trained on the ad title and price.
All the dependencies should be included in theAnaconda distribution of Python 3.4.
nosetests
python ebay_sale_predictor.py --train --keywords "road+bicycle"
python ebay_sale_predictor.py --score --ad-title "vintage mens shimano" --price 900