Noisecut: a python package for noise-tolerant classification of binary data using prior knowledge integration and max-cut solutions

Samadi, Moein E.; Mirzaieazar, Hedieh; Mitsos, Alexander; Schuppert, Andreas

doi:10.1186/s12859-024-05769-8

Table 1 The median with 95% CI of classification metrics for NoiseCut, DNN, XGBoost, SVM, and RF on testing data across different noise intensities in data labeling. The training data size was 70% for all the experiments

From: Noisecut: a python package for noise-tolerant classification of binary data using prior knowledge integration and max-cut solutions

Noise Intensity	Model	Accuracy	Recall	Precision	F1 Score	AUC-ROC
0%	NoiseCut	1.000 ± 0.000	1.000 ± 0.000	1.000 ± 0.000	1.000 ± 0.000	1.000 ± 0.000
	DNN	0.993 ± 0.011	0.998 ± 0.002	0.996 ± 0.006	0.993 ± 0.012	0.999 ± 0.001
	XGBoost	0.974 ± 0.008	0.983 ± 0.019	0.975 ± 0.012	0.975 ± 0.015	0.998 ± 0.004
	SVM	0.934 ± 0.009	0.941 ± 0.015	0.934 ± 0.014	0.929 ± 0.013	0.966 ± 0.009
	RF	0.883 ± 0.010	0.905 ± 0.028	0.880 ± 0.016	0.890 ± 0.021	0.949 ± 0.009
2.5%	NoiseCut	0.974 ± 0.001	0.980 ± 0.006	0.973 ± 0.002	0.975 ± 0.003	0.967 ± 0.003
	DNN	0.933 ± 0.015	0.949 ± 0.034	0.922 ± 0.022	0.934 ± 0.031	0.955 ± 0.012
	XGBoost	0.925 ± 0.008	0.941 ± 0.020	0.927 ± 0.012	0.931 ± 0.016	0.954 ± 0.006
	SVM	0.874 ± 0.009	0.885 ± 0.017	0.879 ± 0.015	0.892 ± 0.015	0.896 ± 0.009
	RF	0.857 ± 0.011	0.873 ± 0.030	0.857 ± 0.017	0.867 ± 0.022	0.908 ± 0.010
5%	NoiseCut	0.947 ± 0.004	0.957 ± 0.013	0.950 ± 0.007	0.951 ± 0.008	0.934 ± 0.004
	DNN	0.891 ± 0.016	0.883 ± 0.046	0.895 ± 0.036	0.892 ± 0.040	0.912 ± 0.019
	XGBoost	0.873 ± 0.008	0.890 ± 0.022	0.880 ± 0.013	0.891 ± 0.017	0.909 ± 0.007
	SVM	0.826 ± 0.011	0.857 ± 0.027	0.832 ± 0.018	0.838 ± 0.025	0.855 ± 0.019
	RF	0.818 ± 0.010	0.844 ± 0.031	0.820 ± 0.015	0.834 ± 0.023	0.865 ± 0.009
7.5%	NoiseCut	0.920 ± 0.005	0.924 ± 0.018	0.921 ± 0.011	0.917 ± 0.012	0.906 ± 0.006
	DNN	0.845 ± 0.019	0.833 ± 0.048	0.828 ± 0.028	0.843 ± 0.032	0.864 ± 0.019
	XGBoost	0.840 ± 0.009	0.868 ± 0.025	0.854 ± 0.014	0.858 ± 0.019	0.875 ± 0.008
	SVM	0.805 ± 0.010	0.819 ± 0.023	0.804 ± 0.016	0.811 ± 0.019	0.824 ± 0.010
	RF	0.798 ± 0.009	0.827 ± 0.033	0.803 ± 0.015	0.810 ± 0.024	0.828 ± 0.010
10%	NoiseCut	0.887 ± 0.006	0.891 ± 0.022	0.892 ± 0.014	0.887 ± 0.017	0.872 ± 0.007
	DNN	0.807 ± 0.016	0.830 ± 0.050	0.803 ± 0.034	0.824 ± 0.041	0.821 ± 0.018
	XGBoost	0.808 ± 0.008	0.831 ± 0.026	0.814 ± 0.014	0.824 ± 0.019	0.839 ± 0.009
	SVM	0.771 ± 0.009	0.786 ± 0.029	0.797 ± 0.016	0.784 ± 0.026	0.798 ± 0.011
	RF	0.771 ± 0.010	0.780 ± 0.034	0.872 ± 0.015	0.773 ± 0.024	0.799 ± 0.011

Back to article page

ISSN: 1471-2105

Contact us

General enquiries: journalsubmissions@springernature.com

BMC Bioinformatics

Contact us