Skip to main content

Table 2 Rules generated from randomForest models using inTrees

From: Transparent computational intelligence models for pharmaceutical tableting process

len

freq

err

Condition

pred

impRRF

2

0.298

0.197

Mg > 0.35 and Dwell ≤ 47.94

Low

1

2

0.205

0.286

Mg ≤ 0.35 and Dwell > 47.94

High

0.44869

2

0.502

0.408

Dwell > 17.75 and Compr. ≤ 24

Medium

0.19967

2

0.059

0.083

SA ≤ 0.405 and MC ≤ 19.795

Low

0.18052

3

0.059

0

Mg ≤ 0.675 and NaCMC > 3.885 and Compr. ≤ 16

Medium

0.16519

4

0.078

0.062

Mg ≤ 0.675 and Dwell > 47.94 and Compr. > 16 and Compr. ≤ 24

High

0.11618

3

0.088

0.111

NaCMC ≤ 3.82 and Dwell ≤ 47.94 and Compr. > 24

Low

0.11316

3

0.059

0

NaCMC ≤ 2.635 and Dwell > 47.94 and Compr. > 24

Medium

0.09774

3

0.088

0

SA ≤ 0.405 and Mg > 0.675 and Dwell ≤ 47.94

Low

0.09270

3

0.059

0

NaCMC ≤ 2.57 and Dwell ≤ 47.94 and Compr. > 24

Low

0.08299

3

0.117

0.167

SA ≤ 1.595 and Mg ≤ 0.35 and Dwell > 47.94

High

0.06932

4

0.088

0.056

SA > 0.575 and Mg > 1.285 and Dwell > 24.52 and Compr. ≤ 24

Medium

0.04240

2

0.307

0.429

SA > 1.595 and Dwell ≤ 72.5

Medium

0.03004

4

0.029

0.167

MC > 25.095 and Mg ≤ 0.35 and Dwell > 24.52 and Dwell ≤ 47.94

Medium

0.02033

1

0.59

0.545

SA ≤ 1.595

Medium

0.01247

3

0.117

0.042

SA ≤ 0.405 and Mg > 0.35 and Dwell ≤ 47.94

Low

0.01165

  1. Mg, MC, SA, NaCMC represent excipients and Dwell and Compr. represents process conditions. Len Length of rules, freq frequency of occurrence of the rule, err error indicating the occurrence of a different prediction, condition: the rule itself, pred prediction if the condition is true, impRRF importance of the rule according to randomForest