LEC. 3: USING R FUNCTIONS

Transcription

1 1 / 19 LEC. 3: USING R FUNCTIONS Instructor: SANG-HOON CHO DEPT. OF STATISTICS AND ACTUARIAL SCIENCES Soongsil University

2 1. Reading Data from (or Writing to) Files 2 / 19 Importing or Exporting Data Large data objects will usually be read as values from external files rather than entered during an R session at the keyboard R input/output facilities are simple but their requirements are fairly strict and even rather inflexible.there is a clear presumption by the designers of R that you will be able to modify your input files using other tools, such as file editors or Perl20 to fit in with the requirements of R. Comma delimited text files For Excel, SPSS, SAS format files, there are many packages available but importing CSV files converted by Excel, SPSS, SAS often simplifies unnecessary complication. Excel files: xlsx SPSS files: foreign, Hmisc SAS files: foreign, Hmisc, sas7bdat

3 1. Reading Data from (or Writing to) Files 3 / 19 Comma Delimited Text Files read.table(), write.table() read.table() function is mainly used for importing data in R read.csv(), write.csv()?read.table write.csv(faithful, "faithful.csv", row.names = FALSE) # if row.anmes = T, index column will be added dir( pattern = "faithful" ) # check "faithful.csv" file created in the working directory dat <- read.table( "faithful.csv", header = TRUE, sep = ",") dat <- read.csv( "faithful.csv") # read.csv is a simple alternative str(dat) data.frame : 272 obs. of 2 variables: $ eruptions: num $ waiting : int

4 1. Reading Data from (or Writing to) Files 4 / 19 Excel Files read.xlsx(), write.xlsx() install.packages("xlsx") --- 현재세션에서사용할 CRAN 미러를선택해주세요 --- URL 을시도합니다 Content type application/x-gzip length bytes (391 KB) ================================================== downloaded 391 KB 다운로드된바이너리패키지들은다음의위치에있습니다 /var/folders/ /1wql8x355w5bd3v25vqdp1q80000gn/T//RtmpCrh3nv/downloaded_packages library(xlsx) 필요한패키지를로딩중입니다 : rjava 필요한패키지를로딩중입니다 : xlsxjars write.xlsx( faithful, file = "faithful.xlsx", sheetname ="testsheet", row.names = FALSE ) dir( pattern = "faithful" ) # check "faithful.xlsx" file created in the working directory [1] "faithful.csv" "faithful.xlsx" dat <- read.xlsx( "faithful.xlsx", sheetname = "testsheet" ) str(dat) data.frame : 272 obs. of 2 variables: $ eruptions: num $ waiting : num

5 2. R Functions for Exploratory Analysis 5 / 19 Summary Statistics sum(), length(), mean(), var(), sd(), median(), max(), min() set.seed(1234) age <- sample( x = 1:100, size = 100, replace = TRUE ) sum(age)/length(age) [1] mean(age) [1] sum((age - mean(age))ˆ2)/(length(age) - 1) [1] var(age) [1] sd(age) [1] median(age) [1] 39 max(age) [1] 100 min(age) [1] 1

6 2. R Functions for Exploratory Analysis 6 / 19 Summary Statistics summary(), range(), IQR(), quantile() summary(age) Min. 1st Qu. Median Mean 3rd Qu. Max range(age) [1] IQR(age) [1] quantile(age) 0% 25% 50% 75% 100% quantile(age, probs = c(.25,.50,.75)) 25% 50% 75%

7 2. R Functions for Exploratory Analysis 7 / 19 Data Manipulation: combing data frames or matrices or vectors cbind(), rbind() The functions cbind and rbind combine data frames, matrices or vectors column-wise and row-wise, respectively. id <- c(3, 1, 5, 10); height <- c(180, 120, 160, 150); weight <- c(90, 22, 80, 40) dat <- cbind(id, height, weight) class(dat) [1] "matrix" cbind(dat, dat) id height weight id height weight [1,] [2,] [3,] [4,] rbind(dat, dat) id height weight [1,] [2,] [3,] [4,] [5,] [6,] [7,] [8,] class(cbind(dat, dat)) [1] "matrix"

8 2. R Functions for Exploratory Analysis 8 / 19 Data Manipulation: sorting sort() The functions sort sorts (or order) a vector or factor (partially) into ascending or descending order dat <- data.frame(dat) class(dat) [1] "data.frame" datt <- rbind(dat, dat) class(datt) [1] "data.frame" datt$height [1] sort(datt$height) [1] sort(datt$height, decreasing = TRUE) [1] rev(sort(datt$height)) # reverse the vector [1]

9 2. R Functions for Exploratory Analysis 9 / 19 Data Manipulation: sorting sort.list() The functions sort.list returns a permutation which rearranges into ascending or descending order ind <- sort.list(datt$height); ind [1] datt$height[ind] [1]

10 2. R Functions for Exploratory Analysis 10 / 19 Data Manipulation: sorting order() Basically, the same as sort.list but order takes more than one argument indd <- order(datt$height); indd [1] datt$height[indd] [1] inddd <- order(datt$id, datt$weight) datt[inddd,] id height weight

11 2. R Functions for Exploratory Analysis 11 / 19 Tables: calculating frequencies table() # the default behavior of read.table is to convert character variables (which are not converted to logical, numeric or complex) to factors dat <- read.csv( "income.csv", colclasses = c("character", "factor", "numeric")) #dat <- read.csv( "income.csv", stringsasfactors = F) str(dat) data.frame : 30 obs. of 3 variables: $ state : chr "tas" "sa" "qld" "nsw"... $ state.fac: Factor w/ 8 levels "act","nsw","nt",..: $ incomes : num # states in Australia attach(dat) table(state.fac) state.fac act nsw nt qld sa tas vic wa income.fac <- factor(cut(incomes, breaks = 35+10*(0:7))) table(incomef,statef) statef incomef act nsw nt qld sa tas vic wa (35,45] (45,55] (55,65] (65,75] detach(dat)

12 2. R Functions for Exploratory Analysis 12 / 19 Tables: calculations on cross-classifications tapply() attach(dat) The following objects are masked _by_.globalenv: incomes, state, state.fac tapply(incomes, state.fac, mean) act nsw nt qld sa tas vic wa round(tapply(incomes, state.fac, mean),2) act nsw nt qld sa tas vic wa detach(dat)

13 2. R Functions for Exploratory Analysis 13 / 19 Tables: calculations on cross-classifications apply() str(faithful) data.frame : 272 obs. of 2 variables: $ eruptions: num $ waiting : num apply( faithful, 2, mean) eruptions waiting apply( faithful, 2, function(x) {sum(x) / length(x)}) eruptions waiting apply( faithful, 2, sd) eruptions waiting

14 3. Probability Distribution Functions 14 / 19 Random Samples and Permutations sample() The function sample can be used to generate random samples with or without replacement #toss a coin 10 times. head = 1, tail = 0 sample(x = c(0,1), size = 10, replace = TRUE, p = c(1/2,1/2)) [1] sample(c(0,1), 10, replace = T) [1] #roll a die 10 times. face values = 1, 2, 3, 4, 5, 6. equally likely sample(x = c(1,2,3,4,5,6), size = 10, replace = TRUE, p = rep(1/6,6)) [1] sample(c(1,2,3,4,5,6), 10, replace = T) [1]

15 3. Probability Distribution Functions 15 / 19 d, p, q, r Functions R has four types of functions for getting information about a family of distributions d(): p.d.f (probability distribution function) p(): c.d.f (cumulative distribution function) q(): quantiles r(): random samples

16 3. Probability Distribution Functions 16 / 19 d, p, q, r Functions Binomial distribution #X: num of successes in n = 10, p = 1/2 Bernoulli trials choose(10,5) * (1/2)ˆ(10) [1] dbinom(x = 5, size = 10, prob = 1/2) [1] pbinom(q = 5, size = 10, prob = 1/2) [1] sum( dbinom(x = 0:5, size = 10, prob = 1/2) ) [1] pbinom(q = 5, size = 10, prob = 1/2, lower.tail = T) [1] pbinom(q = 5, size = 10, prob = 1/2, lower.tail = F) [1] pbinom(q = 5, size = 10, prob = 1/2, lower.tail = T) [1] pbinom(q = 5, size = 11, prob = 1/2) [1] 0.5 qbinom(p = 0.5, size = 11, prob = 1/2) [1] 5

17 3. Probability Distribution Functions 17 / 19 d, p, q, r Functions Binomial distribution prob <- dbinom(0:10, size = 10, prob = 1/2) plot(0:10, prob, type = "h", col = 4, lwd = 4) points(0:10, prob, col = 4, pch = 16) cumprob <- pbinom(0:10, size = 10, prob = 1/2, lower.tail = T) plot(0:10, cumprob, type = "s", col = 4) plot(0:10, cumsum(prob), type = "s", col = 4) dat.bin <- rbinom(n = 1000, size = 10, prob = 1/2) hist(dat.bin) mean(dat.bin) [1] sd(dat.bin) [1]

18 3. Probability Distribution Functions 18 / 19 d, p, q, r Functions Normal distribution #X N(0,1) dnorm(x = 0, mean = 0, sd = 1) [1] pnorm(q = 0, mean = 0, sd = 1) [1] 0.5 qnorm(p = 0.5, mean = 0, sd = 1) [1] 0 dat.norm <- rnorm(n = 1000, mean = 0, sd = 1) hist(dat.norm, freq = F) mean(dat.norm) [1] sd(dat.norm) [1] xx <- seq(min(dat.norm), max(dat.norm), by = 0.1) yy <- dnorm(x = xx, mean = mean(dat.norm), sd = sd(dat.norm)) lines( xx, yy, col = 2, lwd = 3 )

19 19 / Probability Distribution Functions 1.1 Probability Distributions R function Table names 1.1: Function and parameters names and parameters for standard for probability standard probability distributions distributions. Distribution Name Parameters beta beta shape1, shape2 binomial binom size, prob Cauchy cauchy location, scale chi-squared chisq df exponentila exp rate F f df1, df2 gamma gamma shape, rate geometric geom prob hypergeometric hyper m, n, k log-normal lnorm meanlog, sdlog logistic logis location, scale negative binomial nbinom size, prob normal norm mean, sd normal range nrange size Poisson pois lambda stable stab index, skewness T t df uniform unif min, max Weibull weibull shape, scale Wilcoxon wilcox m, n multivariate normal mvnorm mean, cov...