[Retracted] Fine-Tuning Word Embeddings for Hierarchical Representation of Data Using a Corpus and a Knowledge Base for Various Machine Learning Applications

Table 2

Classifier performance using different embedding methods as features on several hypernym benchmark datasets with concatenation as an operator to represent the relation.

Model	BLESS		BARONI		KOTLERMAN		LEVY		WEEDS
		AUC		AUC		AUC		AUC		AUC
CBOW	88.41	87.43	67.84	68.30	53.79	54.72	67.41	67.47	62.27	62.50
SGNS	87.47	86.29	67.66	68.04	56.77	57.11	70.98	68.13	63.21	63.48
GloVe	91.85	93.28	68.87	69.33	57.61	57.72	68.47	69.78	66.54	66.67
R-CBOW	84.43	79.04	68.64	68.76	48.46	52.44	50.03	51.06	66.61	66.83
R-SGNS	83.61	78.08	69.70	70.04	49.84	53.71	48.93	50.51	69.06	69.28
JR	89.86	88.94	68.95	69.48	54.76	55.38	67.60	68.06	66.96	67.12
HyperVec	86.56	82.78	73.82	74.26	54.30	55.51	57.63	57.78	74.65	74.77
LEAR	92.84	93.98	74.63	74.47	57.53	57.24	70.96	75.23	74.98	75.03
Poincaré	66.96	80.61	63.97	64.84	53.49	56.27	52.22	61.85	62.45	62.89
HWE	88.19	90.23	74.72	75.03	55.95	57.55	71.92	76.66	72.17	72.34