It provides a similarity score computation component to pre-compute scores based on WordNet/GermaNet. The PATH and HSO measures are directly available, other scores can be added by extending the code. (The APIs offer a multitude of scores, but we focus our evaluation mainly on the Path measure.)
The other component is the DT evaluation component. Any kind of DT can be evaluated, if it is in the correct format, consisting of three fields:
The fields should be TAB-separated and the entries for each word1 should ordered by score (descending). A word can contain a POS tag: word#NN. For further instructions, consult the documentation page.