test

.gitignore

+.Rproj.user
+.Rhistory
+.RData
+.Ruserdata

Day1-profiling.R

+####Part 1, Profiling (Ron):
+##R packages: ggplot2, ggbeeswarm, bench, Rprof, proftools,profvis, data.table, bench
+
+install.packages("ggbeeswarm")
+install.packages("bench")
+install.packages("Rprof")
+install.packages("proftools")
+install.packages("profvis")
+install.packages("data.table")
+install.packages("bench")
+install.packages("microbenchmark")
+library(ggbeeswarm)
+library(bench)
+library(Rprof)
+library(proftools)
+library(profvis)
+library(data.table)
+library(bench)
+library(ggplot2)
+library(microbenchmark)
+library(ggplot2)
+
+########################################################
+#Task 1: compare the speed of mean(x) with sum(x) / length(x) for different lengths of x.
+x<- rep(1,1000000)
+compare <- microbenchmark(mean(x), sum(x) / length(x), times = 1000)
+compare <- microbenchmark(mean.default(x), mean(x), times = 1000)
+autoplot(compare)
+autoplot(compare)
+
+########################################################
+#Task 2: compare the speed of selecting one column in a matrix and in a data.frame object.
+#Does the difference depend on the object size?
+#with 30000 data
+
+dat_matrix<-matrix(1,10000,3)
+dat_df<-data.frame(dat_matrix)
+compare <- microbenchmark(dat_df[,2], dat_matrix[,2], times = 1000)
+autoplot(compare)
+
+#with 100 data
+dat_matrix<-matrix(1,100,3)
+dat_df<-data.frame(dat_matrix)
+compare <- microbenchmark(dat_df$x.1, dat_matrix[,2], times = 1000)
+autoplot(compare)
+#small data not very difference
+
+#Do the same for the selection of a row. What do you notice?
+#with 30000 data
+x<-rep(1,10000)
+dat_matrix<-matrix(1,3,10000)
+dat_df<-data.frame(rbind(x,x,x))
+compare <- microbenchmark(dat_df[1,], dat_matrix[1,], times = 100)
+autoplot(compare)
+
+##################################################
+### Case study: cross validation
+require(pls)
+example(plsr)
+r1 <- yarn$density - yarn.pls$validation$pred[,1,1]
+r2 <- yarn$density - yarn.pls$validation$pred[,1,2]
+
+#######################################################
+## datatable
+library(data.table)
+flights <-  fread("https://github.com/Rdatatable/data.table/blob/master/vignettes/flights14.csv?raw=TRUE") 
+# flight data is data table
+flight_df<-as.data.frame(flights)
+
+#use the flights data, and compare the speed of selecting rows
+#and columns in both data.table and data.frame formats
+
+
+#select row
+compare <- microbenchmark(flights [1,], flight_df[1,], times = 100)
+autoplot(compare)
+#select column
+compare <- microbenchmark(flights [,.(year,month,day,dep_delay)], flight_df[,1:4], times = 100)
+autoplot(compare)
+
+## dataframe is faster for select one row but for more complex tasks datatable can perform better
+
+##################################################
+#find determinants
+
+a<-as.matrix((0:9))
+b<-c((0:9))
+d<-as.matrix((0:9))
+e<-c((0:9))
+
+ab<-c(a%*%b)
+de<-c(d%*%e)
+
+dd.for.c<-function(){
+  val<-NULL
+  for(a1 in 0:9)
+    for(b1 in 0:9)
+      for(d1 in 0:9)
+        for(e1 in 0:9)
+          val<-c(val,a1*b1-d1*e1)
+        table(val)
+}
+
+a1=((0:9))
+b1=((0:9))
+d1=((0:9))
+e1=((0:9))
+
+dd.for.c(a1=a1,b1=b1,d1=d1,e1=e1)
+
+#improve argument
+dd.for.c2 <- function() {
+  val <- outer(0:9, 0:9, "*")
+  val2 <- outer(c(val), c(val), "-")
+  table(val2)
+}
+
+dd.for.c2 <- function() {
+  val <- outer(0:9, 0:9, "*")
+  val2 <- outer(val, val, "-")
+  tabulate(val2 + 82) # +82 to correct from something from tabulate
+}
+
+?outer #The outer product of the arrays X and Y is the array A with dimension c(dim(X), dim(Y)) 

Day2-bigdata.R

+#Part 2, Big data (Ron):
+# R packages: curl, pryr, bigmemory, biganalytics, biglm
+
+install.packages("curl")
+install.packages("pryr")
+install.packages("bigmemory")
+install.packages("biganalytics")
+install.packages("biglm")
+
+
+library(curl)
+library(pryr)
+library(bigmemory)
+library(biglm)

Day3 machine learning.R

+
+#Part 3 and 4, Parallel computing / Machine learning / research (Sven):
+# R packages: knitr, doParallel, GGally, ggplot2, numbers, OpenCL,
+#parallel, plot3D, randomForest, rgl, Rmpi, rpart, rpart.plot, snow,
+#dplyr, tidyverse, hablar
+
+install.packages("knitr")
+install.packages("doParallel")
+install.packages("GGally")
+install.packages("numbers")
+install.packages("OpenCL")
+install.packages("parallel")
+install.packages("plot3D")
+install.packages("randomForest")
+install.packages("rgl")
+install.packages("Rmpi")
+install.packages("rpart")
+install.packages("rpart.plot")
+install.packages("snow")
+install.packages("hablar")
+install.packages("dplyr")
+install.packages("tidyverse")
+
+library(hablar)
+library(doParallel)
+library(GGally)
+library(numbers)
+library(OpenCL)
+library(parallel)
+library(plot3D)
+library(randomForest)
+library(rgl)
+library(Rmpi)
+library(rpart)
+library(rpart.plot)
+library(snow)
+library(hablar)
+library(dplyr)
+library(tidyverse)

RBigdata.Rproj

+Version: 1.0
+
+RestoreWorkspace: Default
+SaveWorkspace: Default
+AlwaysSaveHistory: Default
+
+EnableCodeIndexing: Yes
+UseSpacesForTab: Yes
+NumSpacesForTab: 2
+Encoding: CP874
+
+RnwWeave: Sweave
+LaTeX: pdfLaTeX