Skip to main content

Table 4 Weka classification parameters

From: Multi-method approach to wellness predictive modeling

Classifier

Parameter name

Parameter value

MLP

hiddenLayers

22

learningRate

0.3

momentum

0.2

trainingTime

500

validationSetSize

0

validationThreshold

20

Rbf network

maxIts

−1

minStdDev

0.1

numClusters

1

ridge

1E−8

J48

confidenceFactor

0.25

minNumObj

2

NB

None

 

BN

estimator

SimpleEstimator

searchAlgorithm

K2

SMO

kernel

PolyKernel

tolerance

0.001

Bagging

classifier

J48

bags

10

bagsize

100 %

AdaboostM1

classifier

NB

iterations

10

RF

trees

2500

features

4