GANN_C {RKEEL}

GANN_C KEEL Classification Algorithm

Description

GANN_C Classification Algorithm from KEEL.

Usage

GANN_C(train, test, hidden_layers, hidden_nodes, transfer, eta,
   alpha, lambda, test_data, validation_data, cross_validation,
   BP_cycles, improve, tipify_inputs, save_all, elite,
   num_individuals, w_range, connectivity, P_bp, P_param,
   P_struct, max_generations, seed)

Arguments

train

Train dataset as a data.frame object

test

Test dataset as a data.frame object

hidden_layers

hidden_layers. Default value = 2

hidden_nodes

hidden_nodes. Default value = 15

transfer

transfer. Default value = "Htan"

eta

eta. Default value = 0.15

alpha

alpha. Default value = 0.1

lambda

lambda. Default value = 0.0

test_data

test_data. Default value = TRUE

validation_data

validation_data. Default value = FALSE

cross_validation

cross_validation. Default value = FALSE

BP_cycles

BP_cycles. Default value = 10000

improve

improve. Default value = 0.01

tipify_inputs

tipify_inputs. Default value = TRUE

save_all

save_all. Default value = FALSE

elite

elite. Default value = 0.1

num_individuals

num_individuals. Default value = 100

w_range

w_range. Default value = 5.0

connectivity

connectivity. Default value = 0.5

P_bp

P_bp. Default value = 0.25

P_param

P_param. Default value = 0.1

P_struct

P_struct. Default value = 0.1

max_generations

max_generations. Default value = 100

seed

Seed for random numbers. If it is not assigned a value, the seed will be a random number

Value

A data.frame with the actual and predicted classes for both train and test datasets.

Examples

data_train <- RKEEL::loadKeelDataset("iris_train")
data_test <- RKEEL::loadKeelDataset("iris_test")

#Create algorithm
algorithm <- RKEEL::GANN_C(data_train, data_test, hidden_layers=1, 
  hidden_nodes=5, max_generations=5)

#Run algorithm
algorithm$run()

#See results
algorithm$testPredictions