Unsupervised Morpheme Analysis -- Morpho Challenge 2010

[an error occurred while processing this directive]

Competition 1 - Alternative results using EMMA

English

**Table:** The submitted morpheme analysis compared to the Gold Standard in **English**. Methods without authors are reference methods by the organizers. The types are: S = semi-supervised algorithm, P = unsupervised algorithm with supervised parameter tuning, U = fully unsupervised algorithm.
AUTHOR	METHOD	PRECISION	RECALL	F-MEASURE	TYPE
Kohonen	Morfessor S+W+L	0.8208	0.8450	0.8327	S
Lignos	Base Inference	0.8744	0.7682	0.8178	U
Lignos	Iterative Compounding	0.8736	0.7641	0.8152	U
Kohonen	Morfessor S+W	0.7939	0.8176	0.8056	S
Lignos	Aggressive Compounding	0.8217	0.7588	0.7890	U
-	Morfessor Baseline	0.8686	0.7226	0.7889	U
-	Morfessor CatMAP	0.9209	0.6815	0.7833	U
Nicolas	MorphAcq	0.7851	0.7759	0.7804	U
Kohonen	Morfessor U+W	0.7313	0.7673	0.7489	P
Spiegler	Promodes-E	0.6573	0.6977	0.6769	P
Spiegler	Promodes	0.5658	0.7382	0.6406	P
Spiegler	DEAP MDL-NOCAT	0.5233	0.8159	0.6376	S
Spiegler	DEAP PROB-NOCAT	0.5618	0.7314	0.6355	S
Golenia	MAGIP	0.4876	0.7083	0.5775	S
Spiegler	Promodes-H	0.4016	0.6298	0.4904	P
Spiegler	DEAP MDL-CAT	0.2579	0.8503	0.3957	S
Spiegler	DEAP PROB-CAT	0.2229	0.7730	0.3460	S
-	Letters	0.0942	0.2953	0.1429	-

Finnish

**Table:** The submitted morpheme analysis compared to the Gold Standard in **Finnish**. Methods without authors are reference methods by the organizers. The types are: S = semi-supervised algorithm, P = unsupervised algorithm with supervised parameter tuning, U = fully unsupervised algorithm.
AUTHOR	METHOD	PRECISION	RECALL	F-MEASURE	TYPE
Kohonen	Morfessor S+W+L	0.6717	0.7573	0.7119	S
Lignos	Base Inference	0.7990	0.5605	0.6588	U
Lignos	Iterative Compounding	0.7733	0.5692	0.6557	U
Lignos	Aggressive Compounding	0.6691	0.6004	0.6329	U
Kohonen	Morfessor S+W	0.5914	0.6533	0.6208	S
-	Morfessor CatMAP	0.7005	0.5425	0.6114	U
-	Morfessor Baseline	0.7614	0.4742	0.5844	U
Kohonen	Morfessor U+W	0.5358	0.5571	0.5462	P
Spiegler	Promodes	0.4493	0.5947	0.5118	P
Spiegler	Promodes-E	0.4423	0.5555	0.4925	P
Golenia	MAGIP	0.4349	0.5364	0.4803	S
Spiegler	DEAP PROB-NOCAT	0.3936	0.5622	0.4629	S
Spiegler	Promodes-H	0.3589	0.5273	0.4271	P
Spiegler	DEAP MDL-NOCAT	0.2960	0.6645	0.4095	S
Spiegler	DEAP MDL-CAT	0.2101	0.6426	0.3166	S
Spiegler	DEAP PROB-CAT	0.2106	0.5095	0.2979	S
-	Letters	0.0929	0.3046	0.1424	-

German

**Table:** The submitted morpheme analysis compared to the Gold Standard in **German**. Methods without authors are reference methods by the organizers. The types are: S = semi-supervised algorithm, P = unsupervised algorithm with supervised parameter tuning, U = fully unsupervised algorithm.
AUTHOR	METHOD	PRECISION	RECALL	F-MEASURE	TYPE
-	Morfessor CatMAP	0.7826	0.5583	0.6517	U
Nicolas	MorphAcq	0.7611	0.5126	0.6125	U
Lignos	Base Inference	0.7498	0.5084	0.6059	U
-	Morfessor Baseline	0.8128	0.4806	0.6040	U
Kohonen	Morfessor U+W	0.6571	0.5560	0.6023	P
Lignos	Iterative Compounding	0.7155	0.5035	0.5910	U
Lignos	Aggressive Compounding	0.6875	0.5117	0.5867	U
-	Letters	0.0834	0.2397	0.1237	-

Turkish

**Table:** The submitted morpheme analysis compared to the Gold Standard in **Turkish**. Methods without authors are reference methods by the organizers. The types are: S = semi-supervised algorithm, P = unsupervised algorithm with supervised parameter tuning, U = fully unsupervised algorithm.
AUTHOR	METHOD	PRECISION	RECALL	F-MEASURE	TYPE
Kohonen	Morfessor S+W+L	0.7532	0.5779	0.6539	S
Kohonen	Morfessor S+W	0.5734	0.4425	0.4994	S
-	Morfessor CatMAP	0.6468	0.4004	0.4946	U
Lignos	Base Inference	0.7131	0.3530	0.4722	U
Nicolas	MorphAcq	0.6310	0.3744	0.4699	U
Lignos	Iterative Compounding	0.6607	0.3602	0.4662	U
Lignos	Aggressive Compounding	0.5709	0.3733	0.4513	U
-	Morfessor Baseline	0.6543	0.3430	0.4500	U
Spiegler	DEAP MDL-NOCAT	0.3590	0.5554	0.4360	S
Spiegler	Promodes	0.4471	0.4117	0.4286	P
Spiegler	DEAP PROB-NOCAT	0.3203	0.5223	0.3971	S
Spiegler	Promodes-H	0.4123	0.3581	0.3833	P
Spiegler	Promodes-E	0.3722	0.3664	0.3693	P
Golenia	MAGIP	0.3479	0.3541	0.3509	S
Spiegler	DEAP MDL-CAT	0.2390	0.5564	0.3344	S
Spiegler	DEAP PROB-CAT	0.2417	0.5206	0.3301	S
Kohonen	Morfessor U+W	0.2959	0.2627	0.2783	P
-	Letters	0.1121	0.2092	0.1459	-

Return to the result page

HOME | RULES | SCHEDULE | DATASETS | EVALUATION | WORKSHOP | FAQ | CONTACT

You are at: CIS → Unsupervised Morpheme Analysis -- Morpho Challenge 2010

Page maintained by webmaster at cis.hut.fi, last updated Tuesday, 24-Aug-2010 17:15:33 EEST