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

Table 6

Accuracy (%) of the different word embedding learning models on the hierarchical path prediction dataset using the COMP and DH as prediction methods on different score functions over the hierarchical paths. The reported results are the average accuracy scores for unigram, bigram, and trigram paths.

Model	Prediction method
	COMP					DH
	Score function
	CBOW	38.12	28.75	43.33	1.04	18.33	48.54	45.42	45.42	1.04	3.04
SGNS	37.08	30.83	37.08	1.04	29.79	42.29	40.21	40.21	1.04	38.12
GloVe	28.75	21.46	27.71	0.0	19.38	46.46	40.21	41.25	1.04	40.21
R-CBOW	42.29	29.79	37.08	1.04	26.67	49.58	39.17	45.42	1.04	3.04
R-SGNS	38.12	30.83	27.71	1.04	28.75	44.38	42.29	39.17	2.08	3.04
JR	29.79	38.12	38.12	1.04	32.92	41.25	44.38	50.62	1.04	41.25
HyperVec	33.54	21.04	21.04	1.04	27.29	47.08	21.04	21.04	1.04	38.75
LEAR	67.29	19.38	22.5	2.08	16.25	78.75	22.5	22.5	0.0	3.04
Poincaré	75.3	65.61	59.85	0.0	48.33	76.21	63.18	60.76	0.0	68.03
HWE	83.82	83.82	82.36	0.30	62.97	84.79	75.03	71.85	0.61	69.39