Description

This work develops a probabilistic approach to quaternary chemical relationship extraction from scientific text. Largely based on extensions to the Snowball Algorithm.

The tool was used to create a vast database of magnetic phase transition temperatures - the first auto-generated database of its kind.

The resulting database was later used by others to discover novel magnetocaloric effects in $HoB_2$. You can read that work here