PERLEX: A Bilingual Persian-English Gold Dataset for Relation Extraction

Table 2

Frequency of relation types in PERLEX dataset.

Relation type	Train	%	Test	%	Total	%	(e1, e2)	(e2, e1)	Reverse order in PERLEX
Cause-Effect	1003	12.54	328	12.07	1331	12.4	478	853	105
Component-Whole	941	11.76	312	11.48	1253	11.7	632	622	191
Content-Container	540	6.75	192	7.07	732	6.8	527	205	23
Entity-Destination	845	10.56	292	10.75	1137	10.6	1135	2	47
Entity-Origin	716	8.95	258	9.50	974	9.1	779	195	168
Instrument-Agency	504	6.30	156	5.74	660	6.2	119	541	58
Member-Collection	690	8.63	233	8.58	923	8.6	110	813	60
Message-Topic	634	7.92	261	9.61	895	8.4	700	195	37
Product-Producer	717	8.96	231	8.50	948	8.8	431	520	158
Other	1410	17.63	454	16.71	1864	17.4	1864	0	194
Total	8000	100	2717	100	10717	100	4911	3946	1041