Deep Learning- and Word Embedding-Based Heterogeneous Classifier Ensembles for Text Classification

Table 2

Classification accuracies of single classifiers on datasets represented with TF-IDF weighting scheme.

Dataset	MVNB	MNB	SVM	RF	CNN
20News-18828	75.89	91.49	91.50	90.36	89.23
20News-19997	63.91	81.29	63.14	78.56	75.56
Mini-news	77.78	82.93	92.40	90.44	91.7
WebKB4	79.30	86.74	91.15	89.75	91.56
1150haber	85.17	95.30	92.52	95.07	94.37
Milliyet	83.19	83.91	92.75	93.05	78.06
Hurriyet	77.14	82.29	81.44	85.11	87.31
Aahaber	82.06	83.26	80.89	88.26	90.19
Average	78.06	85.90	85.72	88.83	87.25
Computation time	1 h 13 m	1 h 45 m	2 h 17 m	3 h 34 m	5 h 23 m