Forecasting Lung Cancer Diagnoses with Deep Learning Data Science Bowl 2017 Technical Report

We describe our contribution to the 2nd place solution for the 2017 Data Science Bowl [8], the largest Kaggle competition to date in terms of prize pool with a $1 million total pool and 2000 competing teams. The goal of the competition is to produce a system that consumes a CT scan and forecasts the probability that the patient in the scan will be diagnosed with lung cancer within a year of the scan. A less technical solution writeup is avilable at the author’s github. [7]