fn.train_amnfis: Function that train a neural network on the training data and...

Description Usage Arguments Value Examples

Description

Function that train a neural network on the training data and gets the correct clusters and weights for making predictions

Usage

1
fn.train_amnfis(df, X, d, formula, n_clusters)

Arguments

df

input dataframe with the training data. Should also contains the class column

X

input numeric matrix which contains the input data without the class column

d

input numeric vector with correct classes for the training data

formula

input formula specifying which are the descriptors and which is the class column in the df. i.e V9~.

n_clusters

the number of clusters to find

Value

object with the best centroids and the best parameters for the network (weights)

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
library(mlbench)
library(dplyr)
library(caret)

data("PimaIndiansDiabetes")

pima_all <- PimaIndiansDiabetes %>% mutate(diabetes = ifelse(diabetes == "pos", 1, 0))

## split the data to get the training set
pima_train_index <- createDataPartition(y = pima_all$diabetes, p = 0.7989, list = FALSE, times = 1)
df.pima_train <- pima_all[pima_train_index, ]

## get all the columns but not the class
mat.pima_train <- as.matrix(df.pima_train[,1:8])
vec.pima_out_train <- df.pima_train$diabetes

## this instruction get the rigth parameters to make predictions on the test set
obj.res_pima <- fn.train_amnfis(df=df.pima_train, X=mat.pima_train, d=vec.pima_out_train, formula=diabetes~., n_clusters=8)

deybvagm/amnfisClassifier documentation built on May 22, 2019, 12:23 p.m.