Table 2 Performance of the classifiers by feature (ALL: all attributes, TAK: title, abstract, keyword, H: herbs, A: affiliation, J: journal, M: MeSH)
From: DisArticle: a web server for SVM-based discrimination of articles on traditional medicine
Precision | Recall | F-Measure | Accuracy | |
|---|---|---|---|---|
TAK | 0.858 | 0.593 | 0.702 | 0.950 |
TAK + M | 0.887 | 0.682 | 0.771 | 0.959 |
TAK + AJM | 0.879 | 0.71 | 0.785 | 0.961 |
TAK + H | 0.957 | 0.862 | 0.907 | 0.982 |
TAK + HA | 0.953 | 0.889 | 0.92 | 0.985 |
TAK + HJ | 0.954 | 0.893 | 0.923 | 0.985 |
TAK + HM | 0.954 | 0.902 | 0.927 | 0.986 |
TAK + HAJ | 0.95 | 0.897 | 0.923 | 0.985 |
TAK + HAM | 0.95 | 0.908 | 0.928 | 0.986 |
TAK + HJM | 0.952 | 0.903 | 0.927 | 0.986 |
ALL-140 | 0.952 | 0.91 | 0.93 | 0.986 |