A Named Entity-Annotated Corpus of 19th Century Classical Commentaries

Matteo Romanello; Sven Najem-Meyer

doi:10.5334/johd.150

Figures & Tables

Table 1

IAA rates computed on a double-annotated sample of the corpus.

SUB-CORPUS	NERC	EL
English	0.83	0.95
French	0.74	0.87
German	0.85	0.81
Avg.	0.81	0.88

Table 2

Basic statistics for the AjMC NE corpus (version 0.4).

LANG.	FOLD	DOCS	TOKENS	MENTIONS
LANG.	FOLD	DOCS	TOKENS	ALL	FINE	NESTED	%NOISY	%NIL
de	Train	76	22,695	1,738	1,738	11	13.81	0.92
	Dev	14	4,701	403	408	2	11.41	0.49
	Test	16	4,845	382	382	0	10.99	1.83
Total		106	32,241	2,528	2,523	13	13.00	0.99
en	Train	60	30,932	1,823	1,823	4	10.97	1.65
	Dev	14	6,506	416	416	0	16.83	1.68
	Test	13	6,052	348	348	0	10.34	2.59
Total		87	43,490	2,587	2,587	4	11.83	1.78
fr	Train	72	24,669	1,621	1,621	9	30.72	0.99
	Dev	17	5,425	391	391	0	36.32	2.56
	Test	15	5,391	360	207	0	27.50	1.45
Total		104	35,487	2,219	2,372	9	31.16	1.31
Grand Total		297	111,216	7,334	7,334	26	18.66	1.36

Table 3

The tagset of annotated entities in the AjMC NE corpus (version 0.4); for each entity type, the total number of mentions in the corpus and some selected examples are provided.

COARSE TAG SET	FINE TAG SET	NB. MENT.	LINKING	EXAMPLES
pers	pers.author	1,212	yes	“Sophocles”, “Euripid.”
	pers.editor	153		“Triclinius”, “Schndw.”
	pers.myth	933		“Tekmessa”, “Ajax”
	pers.other	237		“Musgrave”, “Perikles”
Total (pers)		2,535
work	work.primlit	1,566	yes	“O. T.”, “Iliad”
	work.seclit	82		“Lexicon Sophocleum”, “L. and S.”
	work.fragm	12		“Frag. adesp.”, “fragm.”
	work.journal	1		“the Cambridge Journal of Philology”
	work.other	3
Total (work)		1,664
object	object.manuscr	25	no	“Laurentianus A”
loc	–	109	yes	“Athènes”, “Salamisinsel”
date	–	26	no	“um 770 v. Chr.”, “A.D. 1618”
scope	–	2,975	no	“1340 f.”, “1083”

A Named Entity-Annotated Corpus of 19th Century Classical Commentaries

Figures & Tables

Figure 1

Table 1

Table 2

Table 3

Paradigm

My account