The Discriminants of Long and Short Duration Failures in Fulfillment Sortation Equipment: A Machine Learning Approach

Table 2

Common classifiers in machines and equipment failure modelling.

Classifiers	Hyperparameters	Values
Logistic regression (LR)	max_iter	500
	Classifier penalty	[None, l1, l2, “elastic net”]
	Classifier c	[100, 10, 1.0, 0.1, 0.01]
	Classifier solver	[“Liblinear,” “newton_cg,” “libfgs”]
k-Nearest neighbor (KNN)	Number of neighbors	[1, 22]
	Metric	[“Euclidean,” “manhattan,” “minkowski”]
	Weights	[“Uniform,” “distance”]
Support vector machines (SVM)	Kernels	[“Linear,” “poly,” “rbf,” “sigmoid”]
	Classifier	[0.05, 0.1, 0.5, 0.7, 1]
	Gamma	[0.05, 0.1, 0.5, 0.7, 1]
Decision trees (cart)	Criterion	[“gini”]
	max_depth	[2, 3, 4, 5]
Random forest (RF)	max_features	[1 to 20]
	n_estimators	[10, 100, 1000]
Naïve bayes (GNB)	Cv	[n_splits = 5]
Gradient boosting (GBC)	n_estimators	[1, 2, 4, 8, 16, 32, 64, 100, 200, 300, 500,1000,10000]
	max_depth	[1, 40]
	learning_rate	[1, 0.5, 0.25, 0.1, 0.05, 0.01]