Performance Analysis of an Optimized ANN Model to Predict the Stability of Smart Grid

Table 2

Hyper-parameters for predictive models.

Predictive models	Logistic regression	KNN	Naïve Bayes	Decision tree	SVM	Random forest	XGBoost	Optimized ANN
Hyper-parameters used	✓ Penalty = l2; ✓ Multi_class = auto; ✓ c = 1.0; ✓ max_iter = 100; ✓ tol = 1e − 4	✓ n_neighbour = 3; ✓ Weight = uniform; ✓ algo = auto; ✓ leafsize = 30; ✓ p = 2; ✓ Metric = minkowski	✓ n_class priors = none; ✓ var_smoothening = 1e − 09	✓ Splitter = best; ✓Criterion = entropy; ✓Max_depth = 90; ✓Min_samples_split = 10	✓ Random_state = 1; ✓ Probability = true `	✓ n_estimator = 100; ✓ criterion = gini; ✓ min_sample_split = 10; ✓ max_feature = auto	✓ object = binary_logisticz; ✓ n_estimator = 10; ✓ max_depth = 15	✓ Three hidden layers; ✓ ReLU + Sigmoid activation function; ✓ Adam optimizer