Table 3 Optimal parameters used for the machine learning models.

From: A hybrid residue based sequential encoding mechanism with XGBoost improved ensemble model for identifying 5-hydroxymethylcytosine modifications

Classifier

Parameter

Value

XGBoost

n_estimators

200

Learning rate

0.01

Max depth

20

Min child weight

10

gamma

0.5

Booster

Gbtree

Objective Function

Binary logistics

Col sample by level

0.5, 0.8, 1.0

lambda (reg_alpha)

0.1, 1

alpha (reg_lambda)

0.1, 1

Random state

42

n_estimators

200

RF

n_estimators

200

Bootstrap

True

Random state

42

Criterion

Entropy

Max_features

Auto

Max_depth

20

min_samples_split

9

min_samples_leaf

5

SVM

Gamma

0.001

Kernal

RBF

C

15

Random state

42

KNN

Nearest neighbors

11

Random state

42

Back to article page