The zip file contains four files:

WN-rules-test-head-results.txt
WN-rules-test-tail-results.txt
WN-rules-valid-head-results.txt
WN-rules-valid-tail-results.txt

WN-rules-test... refers to the result of applying the rule-based approach to the test set of the WordNet dataset.
WN-rules-valid... refers to the result of applying the rule-based approach to the validation set of the WordNet dataset.

...-head-results.txt refers the results with respect to results for queries like <?,r,t>
...-tail-results.txt refers the results with respect to results for queries like <h,r,?>

All results are specified in terms of comma separated triples, the last value refers to the confidence score (or probability).

Sometimes some triples are listed annotated with 0.0. This does not make sense, it might be a rounding issue.
However, they should just be treated as triples annotated with 0.0, in the same way as triples not listed.

The files WN-rules files have been generated with the following setting. 


RuleEngineII.UNSEEN_COUNTER_EXAMPLES = 0;
RuleEngineII.MIN_NUM_OF_HITS = 10;
RuleEngineII.TOP_K = 20;
RuleEngineII.PRECISION_OF_APPROXIMATION = 500;
RuleEngineII.SEARCH_SAMPLESIZE  = 500;
RuleEngineII.COMBINATION_RULE = RuleCombinationType.MAXIMUM;
In short: 0|10|20|500|500|MAX 

This results in:

In short: 0|10|20|500|500|MAX 
hits@1    0.945
hits@10   0.958
hits@100  0.970

Some other results:

0|10|20|100|100|MAX
hits@1    0.944
hits@10   0.956
hits@100  0.962


0|10|20|5000|5000|MAX
hits@1    0.945
hits@10   0.959
hits@100  0.970