大家期待已久（？）的實習課練習題終於出爐了。
本周題目
解答
需要用到的附檔：
TA01.xls
TWNstock.dat
一次打包下載：點我
題目中包含四題統計軟體R的基本練習，我會在第一次實習課介紹R的基本操作，在那之前各位有興趣自行摸索的話當然是最好的！
本次習題不需繳交，請各位同學憑良心作答。（啥）
話說，千萬不要太相信解答，因為EverDark這傢伙出完題目就已經呈現半死的狀態，解答可能不太解答，各位如有任何疑惑都歡迎提出來討論。
以上。
不好意思，因為用CwTeX打R的CODE會有不少問題要修正，我發現解答裡面有一些地方的CODE會有錯誤，因為CwTeX把它視為幕後排版的變數了，上課的時候我會另外再說明。以後關於R習題的解答我會用另外的格式發布。
123
2009年9月20日 晚上8:46以下為更正後的R習題參考答案。
我會在第一次實習課詳細講解內容。
## Department of Economics, NTU
## Statistics (I) TA section
## EverDark(c)
## 09/22, 2009
## Problem Set 7
read.csv("TA01.csv",hearder=F) # wrongly include the first row
TA01<-read.csv("TA01.csv",header=T)
dim(TA01) # check the dimension of your data matrix
TA01[1:2,] # check header for all columns
mean(TA01$visits)
var(TA01$visits)
sd(TA01$visits)
var(TA01[,1])^(1/2) # a differnet way to call out "visits"
median(TA01$visits)
sort(TA01$visits) # if you want to find out the mode on your own (normally not the case though)
table(TA01$visits)
max(table(TA01$visits))
## Problem Set 8
X<-c(1:3);X
Y<-seq(6,2,by=-2);Y
Z<-c(1,4,9);Z
cbind(X,Y,Z)->K;K
det(K)
diag(10,3,4) # also check the usage of diag(K)
solve(K,diag(3))
solve(K) # short cut if we are interested in inversing matrix
K%*%K
result<-K
n<-2
for (i in 1:n)
{
result=result%*%K
}
result
K%*%K%*%K # double check
identical(result,K%*%K%*%K) # triple check! (if you wish)
## Problem Set 9
X<-c(1:100);X
Y<-seq(200,2,by=-2);Y
Z<-numeric(100);Z
for (i in 1:100)
{
Z[i]=i^2
}
Z
cbind(X,Y,Z)->K;K
K$X # fail to run, for possible reasons you may check mode(K)
mode(K)
mode(TA01) # it appears that operator $ works for "list" but not "numeric" object
mean(K[,1]) # [] extraction usually works for things in matrix form
?apply # check the useful function apply() on your own
apply(K,2,mean)
apply(K,2,var)
K<-as.vector(K) # twist matrix K into vector K by column
## Problem Set 10
read.table("TWNstock.dat")
TWNstock<-read.table("TWNstock.dat")
TWNstock[1:2,]
TWNstock<-read.table("TWNstock.dat",header=T)
colnames(TWNstock)
colnames(TWNstock)<-c("TW","TSMC","1","2","3")
library(moments)
skewness(TWNstock$TSMC) # mode(TWNstock) turns out to be "list"
kurtosis(TWNstock)
par(mfrow=c(1,2))
hist(TWNstock$TW,pro=T)
curve(dnorm(x,mean(TWNstock$TW),var(TWNstock$TW)),add=T)
hist(TWNstock$TSMC,pro=T)
curve(dnorm(x,mean(TWNstock$TSMC),var(TWNstock$TSMC)),add=T)
123
2009年9月21日 凌晨1:41