Skip to main content

Table 2 Average LRMSE values, ranking (in parenthesis), and regression slope ( β ^ MathType@MTEF@5@5@+=feaagaart1ev2aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xH8viVGI8Gi=hEeeu0xXdbba9frFj0xb9qqpG0dXdb9aspeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGacaGaaiaabeqaaeqabiWaaaGcbaacciGaf8NSdiMbaKaaaaa@2D8B@ ) in equation (5), for the eight methods under investigation (from Simulation II). An asterisk indicates a p-value less than 0.001.

From: Which missing value imputation method to use in expression profiles: a comparative study and two selection schemes

Method

ALI

ALO

BAL

CAU

GAS

GOL

ROS

SP.AFA

SP.ELU

Overall

Slope

LSA

0.487 (1)

0.323 (1)

0.253 (1)

0.348 (2)

0.311 (1)

0.414 (1)

0.492 (1)

0.147 (3)

0.150 (3)

0.325

-0.292*

LLS

0.550 (4)

0.339 (2)

0.254 (2)

0.355 (1)

0.316 (2)

0.421 (2)

0.584 (4)

0.122 (1)

0.117 (1)

0.332

-0.125*

BPCA

0.524 (6)

0.338 (3)

0.254 (3)

0.346 (3)

0.315 (3)

0.420 (3)

0.555 (5)

0.120 (2)

0.115 (2)

0.340

0.007

PLS

0.538 (7)

0.397 (4)

0.316 (4)

0.439 (4)

0.418 (4)

0.473 (4)

0.591 (6)

0.208 (7)

0.220 (6)

0.373

0.330*

KNN.c

0.509 (3)

0.354 (5)

0.280 (5)

0.402 (5)

0.398 (5)

0.439 (5)

0.535 (3)

0.225 (5)

0.267 (5)

0.379

-0.082*

OLS

0.558 (2)

0.343 (6)

0.259 (6)

0.357 (6)

0.322 (6)

0.425 (6)

0.590 (2)

0.232 (6)

0.268 (7)

0.382

-0.149*

KNN.e

0.503 (5)

0.356 (8)

0.286 (8)

0.408 (8)

0.406 (7)

0.448 (8)

0.527 (7)

0.229 (4)

0.273 (4)

0.400

-0.237*

SVD

0.622 (8)

0.385 (7)

0.301 (7)

0.415 (7)

0.446 (8)

0.465 (7)

0.661 (8)

0.279 (8)

0.316 (8)

0.432

0.547*