Classification of premalignant pancreatic cancer mass-spectrometry data using decision tree ensembles

BMC Bioinformatics

Table 5 Classification results using features selected by Wilcoxon rank test.

Algorithm	Accuracy(%)	TP rate	FP rate	TN rate	FN rate	Sensitivity	Specificity	Precision	Fmeasure	RMSE
C4.5	0.6667	0.90	0.63	0.38	0.10	0.90	0.38	0.65	0.75	0.4683
Random Forest	0.7000	0.79	0.41	0.59	0.21	0.79	0.59	0.71	0.74	0.4401
Bagging	0.6667	0.68	0.35	0.65	0.32	0.68	0.65	0.72	0.69	0.4484
Logitboost	0.6833	0.76	0.41	0.59	0.24	0.76	0.59	0.70	0.73	0.4499
Stacking	0.6667	0.93	0.66	0.34	0.07	0.93	0.34	0.64	0.76	0.4639
Adaboost	0.6611	0.76	0.46	0.54	0.24	0.76	0.54	0.68	0.71	0.4805
Multiboost	0.7000	0.73	0.34	0.66	0.27	0.73	0.66	0.74	0.73	0.5187
Logistic	0.6556	0.77	0.49	0.51	0.23	0.77	0.51	0.67	0.71	0.4362
Naivebayes	0.6944	0.70	0.31	0.69	0.30	0.70	0.69	0.77	0.72	0.4969
Bayesnet	0.6778	0.73	0.39	0.61	0.27	0.73	0.61	0.71	0.71	0.5232
Neural Network	0.6778	0.66	0.30	0.70	0.34	0.66	0.70	0.73	0.68	0.4606
RBFnet	0.5944	0.74	0.59	0.41	0.26	0.74	0.41	0.62	0.67	0.4556
SVM	0.6611	0.71	0.40	0.60	0.29	0.71	0.60	0.71	0.70	0.5760

ISSN: 1471-2105