7 * (3 + 2)/22^3sqrt(9)log(10) # logaritmo neperianoexp(log(10))x <- 5a <- c(1, 2, 3, 4, 5, 6, 7, 8, 9)aa1 <- a/9a1a2 <- a^2a2a3 <- sqrt(a)sum(a)mean.a <- mean(a)mean.ahelp.start()help(log)?logls()rm(a)rm(list=ls())ls()# incluir a novamenteinstall.packages("fGarch")m1 <- matrix(c(1,2,4,3,2,5,4,3,6,1,0,1,5,3,2,1),nrow=4)m1m2 <- matrix(c(1,2,4,3,2,5,4,3,6,1,0,1,5,3,2,1),nrow=8)m2det(m1)m1.inv <- solve(m1)m1.invdet(m1.inv)1/det(m1)Y <- matrix(c(1,2,3,4), nrow=4)Ybeta <- solve(m1)%*%Ybetaa.cum <- cumsum(a)a.cumm3 <- apply(m1,2,cumsum)m3m4 <- apply(m1,2,cumprod)m4m5 <- array(1,c(4,4))m5m6 <- diag(4)m6z<-matrix(c(1,2,3,4,5,6), nrow=3)zmean.f.row <- apply(z,1,mean)mean.f.rowmean.f.col <- apply(z,2,mean)mean.f.col

# functiong <- function(x1,x2,x3){ exp(x1)*cos(x2*pi)*sqrt(x3) }g(0,1,1)

f1 <- function(x) {1 + x}

f2 <- function(x) { 1 + x + x^2/2}f3 <- function(x) {1 + x + x^2/2 + x^3/6}x <- seq(from=-1, to=1, by=0.01)y1 <- f1(x)y2 <- f2(x)y3 <- f3(x)y4 <- exp(x)

plot(x,y1,type="l",col="red", ylab=expression(paste(g(x))), main="aproximação da função exponencial por Taylor")lines(x,y2, type="l", col="green")lines(x,y3,type="l", col="magenta")lines(x,y4, type="l", col="blue", lwd=1, lty=2)legend("topleft", col=c("red","green","magenta","blue"), lty=c(1,1,1,2), legend=c("linear","parábola", "cúbia", "exp"))

# números aleatórios de uma N(0,1)par(mfrow=c(1,1))set.seed(877)y <- rnorm(1000)hist(y,30) # linear regressionlibrary(Ecdat)data(Capm)attach(Capm)head(Capm)length(rfood)rfood <- rfood/100rmrf <- rmrf/100

par(mfrow=c(1,1))plot(rmrf,rfood,ylab="Food industry excess return", xlab="Market excess return")

fit <- lm(rfood~rmrf)abline(fit, col="red")summary(fit) anova(fit)vcov(fit)coefficients(fit)confint(fit, level=0.95)residuals(fit)str(fit)fit$coefficientshist(fit$residuals)

# importando dadoslibrary(fImport)Y <- yahooSeries("PBR", from="2005-01-01", to="2016-05-10")head(Y)price <- Y[,"PBR.Close"]library(quantmod)chartSeries(price, theme="white") chartSeries(price, chartTheme('white')) library(timeSeries)ret <- returns(price)length(price)length(ret)par(mfrow=c(2,1))

plot(price, type="l", ylab="price", xlab="data", main="Petrobras")plot(ret,type ="l", ylab="return", xlab="data")

par(mfrow=c(1,1))hist(ret, n=50, xlim=c(-0.7,0.3))

library(fGarch)par(mfrow=c(1,1))qqnorm(ret); qqline(ret,col=2)x<-seq(-.15,.15,by=0.001)hist(ret,50, prob=TRUE, ylim=c(0,30),xlim=c(-0.15,0.15))lines(density(ret, adjust=1), lwd=2, ylim=c(0,60), xlim=c(-0.05,0.05))lines(x,dstd(x, mean=mean(ret), sd=sd(ret), 4), lty=5, lwd=2, col="red")lines(x,dnorm(x,mean=mean(ret),sd=sd(ret)), lty=3, lwd=4, col="blue")legend("topleft",c("KDE","t: df=5", "normal"), lwd=c(2,2,4), lty=c(1,5,3), col=c("black","red","blue"))

# random walk

set.seed(123456)mu <- 0.99e <- rnorm(500)## pure random walkrw.nd <- cumsum(e)## trendtrd <- 1:500## random walk with driftrw.wd <- mu*trd + cumsum(e)## deterministic trend and noisedt <- e + mu*trd## plottingpar(mar=rep(5,4))plot.ts(dt, lty=1, ylab='', xlab='')lines(rw.wd, lty=2)par(new=T)# plot on the right axisplot.ts(rw.nd, lty=3, axes=FALSE)axis(4, pretty(range(rw.nd)))lines(rw.nd, lty=3)legend(10, 18.7, legend=c('det. trend + noise (ls)', 'rw drift (ls)', 'rw (rs)'), lty=c(1, 2, 3))

# or can use the plot code ....

plot(dt,type="l", lty=1, ylab="", xlab="", col="blue")lines(rw.wd, lty=2, col="red")par(new=T)plot(rw.nd,type="l", lty=3, axes=FALSE, ylab="", xlab="time", col="green")axis(4)legend(40, 18.7, legend=c('det. trend + noise','random walk+drift', 'random walk'),lty=c(1, 2, 3), col=c("blue","red","green"))

# plot 3Dlibrary(scatterplot3d)set.seed(124)x <- rnorm(1000); y <- rnorm(1000)

z <- x^2 + y^2par(mfrow = c(1,1), pty='s')scatterplot3d(x,y,z,highlight.3d = T, col.axis = 1, col.grid = 1, pch=20)