A Variable-Clustering-Based Feature Selection to Improve Positive and Negative Discrimination of P53 Protein in Colorectal Cancer Patients

Table 2

Accuracy and Kappa values of the model under different eigenvalues of the P53 dataset are as follows: min, mean, and max.

Metrics	Condition	Feature	Min	Mean	Max
Accuracy	IV_Cluster_LR	5	0.60	0.67	0.77
	IV_LR	5	0.57	0.64	0.74
	IV_Cluster_LR	15	0.62	0.71	0.86
	IV_LR	15	0.57	0.70	0.83
	IV_Cluster_LR	25	0.57	0.73	0.89
	IV_LR	25	0.56	0.69	0.83
	ALL_LR	ALL	0.34	0.49	0.66
Kappa	IV_Cluster_LR	5	0.19	0.34	0.54
	IV_LR	5	0.11	0.28	0.48
	IV_Cluster_LR	15	0.23	0.42	0.71
	IV_LR	15	0.14	0.38	0.66
	IV_Cluster_LR	25	0.14	0.45	0.77
	IV_LR	25	0.10	0.37	0.65
	ALL_LR	ALL	-0.29	-0.01	0.32