Predicting Hacker News Upvotes
Natural Language Processing of Hacker News article headlines
Description
Using Hacker News Headlines (from 2006-2015Q1) which I vectorized into a bag of words model, I attempted to predict the number of upvotes the aritcle would recieve. I implemented a simple supervised, mutli-layer perceptron Neural Network model, which provided good predictive power on an untrained test set.
The dataset was originally compiled by @arnauddri and can be found in the stories.7z file here.