Skip to content

README.md

Latest commit

 

History

History
24 lines (15 loc) · 940 Bytes

README.md

File metadata and controls

24 lines (15 loc) · 940 Bytes

hexgraph

HEX Graph implementation for fine grained entity recognition

To use HEXGRAPH:

  1. Specify your graph in the form of a .hxg file. The .hxg format is as follows:

#E# enumerate all the classes in your graph (1 per line)

#X# Put exclusion relationships in this section by listing any number of classes separated by spaces. For example, if a line in this section is "A B C", exclusions [A,B], [A,C], and [B,C] will be added.

#H# Put hierarchy relationships in this section using the format [Superclass]: [space separated list of subclasses]

  1. Specify your class scores. This is done by simply creating a text file with the class name and corresponding score value all on one line.

2a. Specify the raw sentences you are processing by including them in a corresponding directory. This step is optional.

  1. Specify the paths to the score files, graph files, and an output destination in ResultRunner.java.

Run your files