Skip to Main Content
Table 5.

Values of precision, recall, and f-measure. Bold indicates the best results

ClassifierDescriptionPrec.Rec.F1
TF-IDF TF-IDF 16.7% 24.0% 19.7% 
TF-IDF-M TF-IDF mapped to CSO concepts 40.4% 24.1% 30.1% 
LDA100 LDA with 100 topics 5.9% 11.9% 7.9% 
LDA500 LDA with 500 topics 4.2% 12.5% 6.3% 
LDA1000 LDA with 1,000 topics 3.8% 5.0% 4.3% 
LDA100-M LDA with 100 topics mapped to CSO 9.4% 19.3% 12.6% 
LDA500-M LDA with 500 topics mapped to CSO 9.6% 21.2% 13.2% 
LDA1000-M LDA with 1,000 topics mapped to CSO 12.0% 11.5% 11.7% 
W2V-W W2V on windows of words 41.2% 16.7% 23.8% 
STM Classifier used by STM 80.8% 58.2% 67.6% 
SYN Syntactic module 78.3% 63.8% 70.3% 
SEM Semantic module 70.8% 72.2% 71.5% 
INT Intersection of SYN and SEM 79.3% 59.1% 67.7% 
CSO-C The CSO Classifier 73.0% 75.3% 74.1% 
ClassifierDescriptionPrec.Rec.F1
TF-IDF TF-IDF 16.7% 24.0% 19.7% 
TF-IDF-M TF-IDF mapped to CSO concepts 40.4% 24.1% 30.1% 
LDA100 LDA with 100 topics 5.9% 11.9% 7.9% 
LDA500 LDA with 500 topics 4.2% 12.5% 6.3% 
LDA1000 LDA with 1,000 topics 3.8% 5.0% 4.3% 
LDA100-M LDA with 100 topics mapped to CSO 9.4% 19.3% 12.6% 
LDA500-M LDA with 500 topics mapped to CSO 9.6% 21.2% 13.2% 
LDA1000-M LDA with 1,000 topics mapped to CSO 12.0% 11.5% 11.7% 
W2V-W W2V on windows of words 41.2% 16.7% 23.8% 
STM Classifier used by STM 80.8% 58.2% 67.6% 
SYN Syntactic module 78.3% 63.8% 70.3% 
SEM Semantic module 70.8% 72.2% 71.5% 
INT Intersection of SYN and SEM 79.3% 59.1% 67.7% 
CSO-C The CSO Classifier 73.0% 75.3% 74.1% 
Close Modal

or Create an Account

Close Modal
Close Modal