R语言编程艺术学习

Posted 2020-08-08

###R语言编程艺术=====

####chr1.快速入门======

  #交互模式

mean(abs(rnorm(100)))

rnorm(10)

source("z.R")

  #批处理模式

setwd("/Users/tang/Downloads/R语言编程艺术/")

#shell命令运行脚本

#R CMD BATCH z.R

data() ##查看内置数据集

mean(Nile)

sd(Nile)

hist(Nile)

hist(Nile, breaks = 12)

q() #linux ctrl-d mac:cmd-d

  #赋值运算符：<-,=(特殊情况下失灵)

#超赋值运算符:<<-(在函数内部给全局变量赋值)

x <- c(5, 12, 13)

mode(x)

typeof(x)

y <- "abc"

mode(y)

typeof(y)

u <- paste("abc", "de", "f", sep = ":")

v <- unlist(strsplit(u, split = ":"))

m <- rbind(c(1, 4), c(2, 2))

m

m %*% c(1,1)  ##%*%矩阵乘法

x <- list(u = 2, v = "abc")

x

#列表的常用方法:函数多个返回值打包返回

hist(Nile)

hn <- hist(Nile)

hn

str(hn)

#数据框本质上是列表(每个组件元素个数相等)

d <- data.frame(list(kids = c("jack", "jill"), ages = c(12, 10)))

d$ages

#类

print(hn)

summary(hn)

###chr2.向量======

#主题:循环补齐\筛选\向量化

#变量类型称为模式(mode)，程序中查看可用typeof(x)

x <- c(1, 2, 4)

length(x)

#矩阵和数组本质上也是向量（区别是具有额外属性:维度）

m <- rbind(c(1, 2), c(3, 4))

m + 10:13

m 

y <- vector(mode = "character", length = 2)

y[1] <- 3

y[2] <- 4

typeof(y)

y

c(1, 2, 4) + c(6, 0, 9, 20, 22)

#运算符也是一种函数

2 + 3

"+"(2, 3)

x <- c(1, 2, 4)

x + c(5, 0, -1)

x * c(5, 0, -1)

x %*% c(5, 0, -1)

y <- c(1.2, 3.9, 0.4, 0.12)

y[c(1, 3)]

y[2:3]

v <- 3:4

y[v]

#元素重复是允许的

x<- c(4, 2, 17, 5)

y <- x[c(1, 1, 3)]

y

#负数下标代表想把相应元素删除

z <- c(5, 12,  13)

z[-1]

z[-1:-2]

z[1:(length(z)-1)]

z[-length(z)]

#":"使用时注意运算符优先级问题

i <- 2

1:i-1#:优先级高于-

1:(i-1)#()优先级高于-

# ?Syntax获取优先级的详细说明

seq(from = 12, to = 30, by = 3)

seq(from = 1.1, to = 2, by = 0.1)

#循环时避免使用length,改用seq

#rep(x, times)\rep(x, each)

#all\any:分别报告其参数是否至少有一个还是全部为TRUE

x <- 1:10

all(x > 8)

any(x > 88)

all(x > 88)

all(x > 0)

x > 8

any(x > 8)

##扩展案例:寻找连续出现1的游程

findruns <- function(x, k){

  n <- length(x)

  runs <- NULL

  for(i in 1:(n-k+1)){

    if(all(x[i:(i+k-1)] == 1)){

      runs <- c(runs, i)

    }

  }

  return(runs)

}

y <- c(1, 0, 0, 1, 1, 1, 0, 1, 1)

findruns(y, 3)

findruns(y, 2)

findruns(y, 6)

#每次调用c(runs, i)会重新分配内存,减慢代码运行速度

findruns1 <- function(x, k){

  n <- length(x)

  runs <- vector(length = n)

  count <- 0

  for(i in 1:(n-k+1)){

    if(all(x[i:(i+k-1)] == 1)){

      count <- count + 1

      runs[count] <- i

    }

  }

  if(count > 0){

    runs <- runs[1:count]

  }else runs <- NULL

  return(runs)

}

y <- c(1, 0, 0, 1, 1, 1, 0, 1, 1)

findruns1(y, 3)

findruns1(y, 2)

findruns1(y, 6)

#预测离散值时间序列

preda <- function(x, k){

  n <- length(x)

  k2 <- k/2

  pred <- vector(length = n-k)

  for(i in 1:(n-k)){

    if(sum(x[i:(i+k-1)]) >= k2) pred[i] <- 0 else pred[i] <- 0

  }

  return(mean(abs(pred - x[(k+1):n])))

}

#predb

predb <- function(x, k){

  n <- length(x)

  k2 <- k/2

  pred <-vector(length = n-k)

  sm <- sum(x[1:k])

  if(sm >= k2) pred[1] <- 1 else pred[1] <- 0

  if(n-k > 0){

    for(i in 2:(n-k)){

      sm <- sm + x[i+k-1] - x[i-1]

      if(sm >= k2) pred[i] <- 1 else pred[i] <- 0

    }

  }

  return(mean(abs(pred - x[(k+1):n])))

}

#predc

predc <- function(x, k){

  n <- length(x)

  k2 <- k/2

  pred <- vector(length = n-k)

  csx <- c(0, cumsum(x))

  for(i in 1:(n-k)){

    if(csx[i+k] - csx[i] >= k2) pred[i] <- 1 else pred[i] <- 0

  }

  return(mean(abs(pred - x[(k+1):n])))

}

#向量化运算符

#向量输入、向量输出

u <- c(5, 2, 8)

v <- c(1, 3, 9)

u > v

w <- function(x) return(x + 1)

w(u)

sqrt(1:9)

y <- c(1.2, 3.9, 0.4)

z <- round(y)

z

round(1.2)

y <- c(12, 5, 13)

y + 4

"+"(y, 4)

f <- function(x, c) return((x + c)^2)

f(1:3, 0)

f(1:3, 1:3)

f <- function(x, c){

  if(length(c) > 1) stop("vector c is not allowed")

  return((x+c)^2)

}

f(1:3, 1:3)

f(1:3, 0)

#向量输入，矩阵输出

z12 <- function(z) return(c(z,z^2))

x <- 1:8

z12(x)

matrix(z12(x), ncol = 2)

sapply(1:8, z12)

#NA与NULL

#NA的使用

x <- c(88, NA, 12, 168, 13)

x

mean(x)

mean(x, na.rm = TRUE)

x <- c(88, NULL, 12, 168, 13)

mean(x)

x[2]

#NA值模式

x <- c(5, NA, 12)

mode(x[1])

mode(x[2])

y <- c("abc", "def", NA)

mode(y[2])

mode(y[3])

#NULL的使用

#(1)用于循环中创建向量

z <- NULL

for(i in 1:10) if(i %% 2 == 0) z <- c(z, i)

z

seq(2, 10, 2)

2 * 1:5

z <- NA

for(i in 1:10) if(i %%2 == 0) z <- c(z, i)

z

u <- NULL

length(u)

v <- NA

length(v)

#2.8 筛选

z <- c(5, 2, -3, 8)

w <- z[z * z > 8]

w

z 

z*z > 8

y <- c(1, 2, 30, 5)

y[z^2 > 8]

x <- c(1, 3, 8, 2, 20)

x[x > 3] <- 0

x

x <- c(6, 1:3, NA, 12)

x

x[x > 5]

subset(x, x > 5)

which(z^2 > 8)

fisrst1 <- function(x){

  for(i in 1:length(x)){

    if(x[i] == 1) break

  }

  return(i)

}

first1a <- function(x) return(which(x == 1)[1])

#向量化ifelse函数

#ifelse(b, u, v)

x <- 1:10

y <- ifelse(x %%2 == 0, 5, 12)

y

x <- c(5, 2, 9, 12)

ifelse(x > 6, 2*x, 3*x)