This repository contains code to replicate the experiments in the ICWSM'17 paper entitled "Identifying Leading Indicators of Product Recalls from Online Reviews using Positive Unlabeled Learning and Domain Adaptation," by Shreesh Kumara Bhat and Aron Culotta. A full version of the paper is here: https://arxiv.org/abs/1703.00518
The code was developed in a Jupyter Notebook running Python 3.4.3.
To get started, first install python dependencies:
pip install -r requirements.txt
jupyter notebook Replication.ipynb
You can then click
Cell->Run All to reproduce all the figures and tables in the paper.