In AlfonsoRReyes/RepDataPeerAssessment1: Reproducible Research Assignment 1. Developed as a package

Datasets

iris

summary(iris)
dim(iris)

library(RepDataPeerAssessment1)
library(rprojroot)

project.data <- find_package_root_file('data')
project.extdata <- find_package_root_file('inst/extdata')
project.R <- find_package_root_file('R')
project.data
project.R
project.extdata

Read ecoli dataset

Source: https://www.r-bloggers.com/classification-trees/

ecoli <- read.table(paste(project.extdata, "ecoli.data", sep = "/"))
ecoli

names(ecoli) <- c("Sequence", "mcv", "gvh", "lip", "chg", "aac", "alm1", "alm2", "class")
ecoli

save(ecoli, file = paste(project.extdata, "ecoli.rda", sep = "/"))

xtabs(~ class, data = ecoli)

Step 1: Introduce MCAR NAs in the iris dataset

Missing Completely at Random filling using the uniform distribution.

set.seed(971)
y <- iris$Sepal.Length
ni <- 150
nj <- 1
prop.m <- 0.05
mcar <- runif(ni*nj, min = 0, max = 1)
y.mcar <- ifelse(mcar < prop.m, NA, y)
y.mcar
mean(is.na(y.mcar))

iris$Sepal.Length <- y.mcar

mean(is.na(iris$Sepal.Length))
summary(iris)

# find percentage NAs in all data frame columns
colMeans(is.na(iris))

library(VIM)
aggr_plot <- aggr(iris, col=c('navyblue','red'), numbers=TRUE, sortVars=TRUE, labels=names(data), cex.axis=.7, gap=3, ylab=c("Histogram of missing data","Pattern"))

library(mice)
library(VIM)

md.pattern(iris)

md.pairs(iris)

marginplot(iris[c(1,2)])

imp1 <- mice(iris)

imp1

imp1$imp$Sepal.Length

# appending multiple imputations datasets to original dataset
# .imp variable contains the number of dataset with `0` being the original
# there will be 150 * (5+1) = 900 rows
#
imp_tot2 <- complete(imp1, 'long', inc=TRUE)
imp_tot2

##labels observed data in blue and imputed data in red for y1
col<-rep(c("blue", "red")[1+as.numeric(is.na(imp1$data$Sepal.Length))],6)

##plots data for y1 by imputation
stripplot(Sepal.Length~.imp, data=imp_tot2, jit=TRUE,col=col, xlab="imputation Number")

library(ggplot2)
library(lattice)

AlfonsoRReyes/RepDataPeerAssessment1 documentation built on May 5, 2019, 4:53 a.m.