Recall Network: A Simple Brain-Inspired Algorithm for Classification

Table 5

Performance of all models among 263 datasets.

		Average accuracy (%)	Kappa statistic	Root mean squared error	Mean absolute error	Relative absolute error (%)	Root relative squared error (%)	Total running time (i7, 16G)
RN		73.57	0.42	0.33	0.27	79.08	83.25	3,780s
NB		72.05	0.50	0.34	0.19	60.74	85.51	60s
J48		78.53	0.59	0.28	0.16	44.76	70.59	180s
KNN		76.87	0.56	0.32	0.16	45.24	80.47	780s
SVM1		71.96	0.42	0.36	0.17	49.04	90.03	8,100s
SVM2		72.66	0.44	0.36	0.17	47.95	88.26	8,100s
SVM3		73.01	0.45	0.35	0.16	47.48	87.19	8,100s
ANN1		77.95	0.59	0.28	0.15	43.61	69.93	34.080s
ANN2		77.84	0.59	0.28	0.15	43.27	70.66	34.080s
OneR		67.16	0.40	0.39	0.19	56.74	98.86	60s
Friedman ranked test	Q (95% threshold: 21.67)	218.45	320.03	467.68	555.13	575.90	460.46	—
ANOVA	F (95% threshold: 1.88)	8.20	15.05	13.13	16.48	29.35	18.82	—