. Author manuscript; available in PMC: 2017 Feb 1.

Published in final edited form as: IEEE Trans Pattern Anal Mach Intell. 2015 Sep 23;38(8):1640–1650. doi: 10.1109/TPAMI.2015.2481404

TABLE 3.

Classification accuracies on the DGS dataset for the proposed approach as well as several state-of-the-art methods. SPBoost/Tree: Sequential Pattern Boosting/Tree. RW: Random Walk.

Method	Accuracy	Time/Sample
Markov Chain [40]	50.6% ± 7.1%	–
SPBoost [39]	54.6% ± 8.2%	–
SPTree [40]	55.4% ± 8.4%	–
RW Kernel + SVM [24]	68.87% ± 8.60%	3.86 s
Path Kernel + RKLR (l₁)	63.72% ± 10.26%	–
Path Kernel + RKLR (l₂)	66.97% ± 9.35%	–
Path Kernel + SVM	70.17% ± 8.3%	2.68 s