Skip to main content

Table 1 Overall results of different clustering methods

From: A comparative study of clustering methods on gene expression data for lung cancer prognosis

  

TCGA-LUAD

Shedden

Type

Method

Mean p-value

Standard Deviation

95% CI

Mean p-value

Standard Deviation

95% CI

Unsupervised

KM

0.493

0.107

(4.32E-09, 0.516)

3.06E-02

8.46E-02

(2.25E-14, 0.244)

 

GM

0.355

0.293

(4.32E-09, 0.857)

0.308

0.260

(2.25E-14, 0.892)

 

AC

1.23E-08

0

(1.23E-08, 1.23E-08)

0.276

5.55E-17

(0.276, 0.276)

 

Consensus

4.69E-02

0.167

(1.23E-08, 0.648)

0.408

0.404

(1.19E-05, 0.876)

Semi-Supervised

RF-KM

4.32E-09

1.42E-23

(4.32E-09, 4.32E-09)

2.25E-14

6.83E-28

(2.25E-14, 2.25E-14)

 

RF-GM

4.32E-09

1.55E-23

(4.32E-09, 4.32E-09)

2.25E-14

6.83E-28

(2.25E-14, 2.25E-14)

 

RF-AC

4.32E-09

0

(4.32E-09, 4.32E-09)

2.25E-14

0

(2.25E-14, 2.25E-14)

 

RF-Consensus

4.32E-09

1.85E-11

(4.32E-09, 4.32E-09)

2.25E-14

6.88E-28

(2.25E-14, 2.25E-14)

 

Cox-KM

4.32E-09

1.48E-23

(4.32E-09, 4.32E-09)

2.25E-14

6.79E-28

(2.25E-14, 2.25E-14)

 

Cox-GM

4.32E-09

1.56E-23

(4.32E-09, 4.32E-09)

1.14E-03

1.15E-02

(2.25E-14, 2.25E-14)

 

Cox-AC

4.32E-09

0

(4.32E-09, 4.32E-09)

2.25E-14

0

(2.25E-14, 2.25E-14)

 

Cox-Consensus

4.32E-09

1.53E-23

(4.32E-09, 4.32E-09)

2.25E-14

6.91E-28

(2.25E-14, 2.25E-14)

Supervised

SC

4.32E-09

2.71E-22

(4.32E-09, 4.32E-09)

2.25E-14

1.24E-27

(2.25E-14, 2.25E-14)

  1. Note: This table displays the overall clustering results, providing p-values, standard deviations for p-values, and 95% confidence intervals (CIs). Methods are organized by type (unsupervised, semi-supervised, and supervised). Results are summarized from 200 trials for each method. Agglomerative clustering results, except for RF-AC, do not have a standard deviation or confidence interval, as they are not affected by random initial values and thus are the same across 200 trials. The 95% CI was calculated based on the 2.5 and 97.5 percentile of p-values from 200 trials
Back to article page

BMC Research Notes

ISSN: 1756-0500

Contact us