Notes

I was inspired by this post on HN, and decided that I could build my own.

This is a pretty rough first effort.

The project consists of 3 modules

jarchive - Web scraper for J-Archive. Extracts database of historical jeopardy clues/answers.
jtextclust - Text clustering routines in R to cluster similar questions.
jtutor - Django App to quiz players. Designed to be browser/smartphone compatible. Select questions based on learned history (optimized to get player to profitability very quickly).

Future enhancements may include

Player simulators.
Profitability estimators.
Per-player tracking.
Better optimization model.