Skip to main content

Table 2 Hyperparameters of the model

From: Design and implementation of I/O performance prediction scheme on HPC systems through large-scale log analysis

Name

Description

Linear

None

Polynomial

degree=2

KNN

n_neighbors=2

RF

random_state_42, n_estimators=1000

GBDT

loss=ls, n_estimators=1000, max_depth=3, subsample=1.0, random_state=42

MLP

2 layers, num_nerons=256, activation=relu, loss=mse, optimizer=adam

CNN

num_filters=256, size_filter=3X3, activation=relu, inputshape=X_test.columns, loss=mse, optimizer=RMSprop(0.01)