Table 5.

Cumulative ratio of synonym prediction in each word embedding model. (FT: fastText, W2V:Word2vec, SKIP:skip-gram, number: vector dimensions).

Model	Precition	Recall	F1-score	Accuracy
FT_CBOW_100	0.5286	0.8819	0.6610	0.4937
FT_CBOW_200	0.5557	0.8870	0.6833	0.5190
FT_CBOW_300	0.5567	0.8872	0.6841	0.5199
FT_CBOW_400	0.5515	0.8863	0.6799	0.5151
FT_CBOW_500	0.5484	0.8857	0.6774	0.5121
FT_CBOW_600	0.5432	0.8847	0.6731	0.5073
FT_CBOW_700	0.5359	0.8834	0.6671	0.5005
FT_CBOW_800	0.5369	0.8836	0.6680	0.5015
FT_CBOW_900	0.5276	0.8817	0.6602	0.4927
FT_SKIP_100	0.4381	0.8609	0.5807	0.4091
FT_SKIP_200	0.4984	0.8757	0.6353	0.4655
FT_SKIP_300	0.4984	0.8757	0.6353	0.4655
FT_SKIP_400	0.5213	0.8805	0.6549	0.4869
FT_SKIP_500	0.5130	0.8788	0.6478	0.4791
FT_SKIP_600	0.5317	0.8826	0.6636	0.4966
FT_SKIP_700	0.5369	0.8836	0.6680	0.5015
FT_SKIP_800	0.5297	0.8821	0.6619	0.4947
FT_SKIP_900	0.5317	0.8826	0.6636	0.4966
W2V_CBOW_100	0.3101	0.8142	0.4491	0.2896
W2V_CBOW_200	0.3757	0.8415	0.5194	0.3508
W2V_CBOW_300	0.3809	0.8433	0.5247	0.3557
W2V_CBOW_400	0.3954	0.8482	0.5394	0.3693
W2V_CBOW_500	0.3985	0.8492	0.5425	0.3722
W2V_CBOW_600	0.3913	0.8468	0.5352	0.3654
W2V_CBOW_700	0.3965	0.8486	0.5404	0.3703
W2V_CBOW_800	0.4089	0.8525	0.5527	0.3819
W2V_CBOW_900	0.3944	0.8479	0.5384	0.3683
W2V_SKIP_100	0.3195	0.8187	0.4596	0.2983
W2V_SKIP_200	0.3278	0.8225	0.4688	0.3061
W2V_SKIP_300	0.3195	0.8187	0.4596	0.2983
W2V_SKIP_400	0.3195	0.8187	0.4596	0.2983
W2V_SKIP_500	0.3070	0.8127	0.4456	0.2867
W2V_SKIP_600	0.2955	0.8068	0.4326	0.2760
W2V_SKIP_700	0.2872	0.8023	0.4230	0.2682
W2V_SKIP_800	0.2893	0.8035	0.4254	0.2702
W2V_SKIP_900	0.2851	0.8012	0.4206	0.2663

The underlined (bold) numbers show the highest cumulated accuracy in the different models.