A clustering-based topic model using word networks and word embeddings

Mu, Wenchuan; Lim, Kwan Hui; Liu, Junhua; Karunasekera, Shanika; Falzon, Lucia; Harwood, Aaron

doi:10.1186/s40537-022-00585-4

Journal of Big Data

Table 7 Comparison of ClusTop algorithm against various baselines, in terms of Precision (Pre), Recall (Rec) and F-score (FS) for the top 10 keywords/unigrams of each topic

From: A clustering-based topic model using word networks and word embeddings

Algorithm	Top 10 Keywords/Unigrams
	Dataset A			Dataset B			Dataset C			Average
	Precision	Recall	F-score	Precision	Recall	F-score	Precision	Recall	F-score	Rank@10
ClusTop-Word-NA	.690 ± .002 (24)	.033 ± .000 (7)	.062 ± .000 (7)	.804 ± .003 (20)	.051 ± .000 (3)	.095 ± .001 (3)	.764 ± .010 (22)	.033 ± .001 (6)	.062 ± .001 (6)	(10.9)
ClusTop-BiG-NA	.707 ± .002 (22)	.035 ± .000 (2)	.065 ± .000 (2)	.789 ± .004 (22)	.052 ± .000 (2)	.096 ± .001 (2)	.765 ± .010 (20)	.035 ± .001 (3)	.067 ± .002 (2)	(8.6)
ClusTop-TriG-NA	.717 ± .002 (21)	.034 ± .000 (4)	.064 ± .000 (4)	.811 ± .003 (19)	.053 ± .000 (1)	.098 ± .001 (1)	.746 ± .010 (23)	.036 ± .001 (1)	.068 ± .001 (1)	(8.3)
ClusTop-BiHa-NA	.719 ± .002 (20)	.034 ± .000 (4)	.064 ± .000 (4)	.782 ± .004 (24)	.051 ± .000 (3)	.095 ± .001 (3)	.765 ± .010 (20)	.035 ± .001 (3)	.065 ± .001 (3)	(9.3)
ClusTop-Hash-NA	.860 ± .004 (8)	.023 ± .000 (24)	.045 ± .000 (17)	.896 ± .008 (9)	.032 ± .001 (17)	.061 ± .001 (17)	.925 ± .012 (8)	.022 ± .000 (18)	.043 ± .001 (18)	(15.1)
ClusTop-Noun-NA	.736 ± .002 (16)	.033 ± .000 (7)	.062 ± .000 (7)	.803 ± .003 (21)	.049 ± .000 (5)	.091 ± .001 (5)	.802 ± .009 (16)	.035 ± .001 (3)	.065 ± .001 (3)	(9.2)
ClusTop-H2VG-NA	.904 ± .004 (3)	.024 ± .000 (21)	.046 ± .000 (16)	.973 ± .004 (1)	.032 ± .001 (17)	.061 ± .001 (17)	.964 ± .008 (3)	.023 ± .001 (16)	.044 ± .001 (16)	(12.2)
ClusTop-H2VW-NA	.894 ± .001 (5)	.026 ± .001 (17)	.026 ± .001 (22)	.944 ± .001 (4)	.029 ± .001 (20)	.029 ± .001 (20)	.983 ± .001 (1)	.020 ± .001 (23)	.020 ± .001 (23)	(15.0)
ClusTop-H2VF-NA	.892 ± .001 (6)	.026 ± .001 (17)	.026 ± .001 (22)	.931 ± .001 (8)	.029 ± .001 (20)	.029 ± .001 (20)	.943 ± .001 (7)	.021 ± .001 (20)	.021 ± .001 (20)	(15.6)
ClusTop-Word-AH	.692 ± .003 (23)	.025 ± .000 (19)	.049 ± .000 (14)	.784 ± .004 (23)	.043 ± .000 (8)	.080 ± .001 (8)	.778 ± .010 (18)	.030 ± .001 (8)	.057 ± .001 (8)	(14.3)
ClusTop-Hash-AH	.833 ± .002 (11)	.027 ± .000 (12)	.051 ± .000 (13)	.842 ± .004 (15)	.041 ± .000 (11)	.078 ± .001 (11)	.857 ± .009 (12)	.025 ± .000 (13)	.048 ± .001 (13)	(12.3)
ClusTop-Noun-AH	.735 ± .002 (17)	.024 ± .000 (21)	.045 ± .000 (17)	.827 ± .004 (17)	.041 ± .000 (11)	.076 ± .001 (14)	.828 ± .009 (13)	.031 ± .001 (7)	.059 ± .001 (7)	(13.8)
ClusTop-H2VG-AH	.824 ± .002 (12)	.027 ± .000 (12)	.053 ± .000 (12)	.850 ± .003 (12)	.043 ± .000 (8)	.080 ± .001 (8)	.869 ± .009 (11)	.028 ± .001 (9)	.053 ± .001 (9)	(10.3)
ClusTop-H2VW-AH	.908 ± .001 (1)	.024 ± .001 (21)	.024 ± .001 (24)	.944 ± .001 (4)	.027 ± .001 (23)	.027 ± .001 (23)	.964 ± .001 (4)	.020 ± .001 (23)	.020 ± .001 (23)	(16.2)
ClusTop-H2VF-AH	.905 ± .001 (2)	.027 ± .001 (15)	.027 ± .001 (20)	.934 ± .001 (6)	.029 ± .001 (20)	.029 ± .001 (20)	.949 ± .001 (5)	.021 ± .001 (20)	.021 ± .001 (20)	(14.2)
ClusTop-Word-AM	.734 ± .001 (18)	.036 ± .000 (1)	.068 ± .000 (1)	.828 ± .004 (16)	.039 ± .000 (16)	.073 ± .001 (16)	.709 ± .013 (24)	.023 ± .000 (16)	.044 ± .001 (16)	(13.8)
ClusTop-Hash-AM	.731 ± .002 (19)	.030 ± .000 (10)	.058 ± .000 (10)	.845 ± .004 (14)	.043 ± .000 (8)	.081 ± .001 (7)	.823 ± .011 (14)	.028 ± .001 (9)	.053 ± .001 (9)	(11.1)
ClusTop-Noun-AM	.860 ± .001 (8)	.024 ± .000 (21)	.047 ± .000 (15)	.932 ± .003 (7)	.040 ± .000 (14)	.076 ± .001 (14)	.908 ± .008 (9)	.024 ± .000 (14)	.047 ± .001 (14)	(12.9)
ClusTop-H2VG-AM	.800 ± .002 (13)	.031 ± .000 (9)	.060 ± .000 (9)	.875 ± .003 (11)	.041 ± .000 (11)	.078 ± .001 (11)	.882 ± .008 (10)	.027 ± .000 (11)	.051 ± .001 (11)	(10.7)
ClusTop-H2VW-AM	.904 ± .001 (3)	.027 ± .001 (15)	.027 ± .001 (20)	.964 ± .001 (2)	.028 ± .001 (22)	.028 ± .001 (22)	.972 ± .001 (2)	.021 ± .001 (20)	.021 ± .001 (20)	(14.0)
ClusTop-H2VF-AM	.891 ± .001 (7)	.027 ± .001 (15)	.027 ± .001 (20)	.951 ± .001 (3)	.027 ± .001 (23)	.027 ± .001 (23)	.944 ± .001 (6)	.020 ± .001 (23)	.020 ± .001 (23)	(15.9)
LDA-Orig	.848 ± .002 (10)	.029 ± .000 (11)	.056 ± .000 (11)	.885 ± .003 (10)	.040 ± .000 (14)	.076 ± .001 (14)	.808 ± .011 (15)	.026 ± .001 (12)	.051 ± .001 (11)	(12.0)
LDA-Hash	.759 ± .002 (14)	.034 ± .000 (4)	.064 ± .000 (4)	.847 ± .003 (13)	.041 ± .000 (11)	.078 ± .001 (11)	.778 ± .010 (18)	.034 ± .001 (5)	.064 ± .001 (5)	(9.4)
LDA-Ment	.752 ± .002 (15)	.033 ± .000 (7)	.063 ± .000 (6)	.820 ± .004 (18)	.044 ± .000 (6)	.082 ± .001 (6)	.787 ± .013 (17)	.024 ± .000 (14)	.047 ± .001 (14)	(11.4)

The rank of an algorithm’s performance for each metric are provided in brackets

Back to article page