R Programming # 6

# 1. Draw the surface plot of bivariate gumbel copula with parameter

# equals to 1.5 and bivariate normal copula with parameter equals to 0.7.

surface=function()

{

library("copula");

gc=gumbelCopula(1.5, dim=2)

persp(gc, dcopula, col="red")

nc=normalCopula(0.7)

persp(nc, dcopula, col="blue")

}

=============================

# 2. Draw the contour plot of the density and cdf of bivariate gumbel

#copula with with parameter 1.4 and marginals N(3,4) and T(3).

cont=function()

{

library("copula");

#Density contour Plot

density = mvdc(copula = archmCopula(family = "gumbel", param = 1.4), margins = c("norm", "t"), paramMargins = list(list(mean = 0, sd = 1), list(df=3, ncp = 3)))

print(density)

contour(density, dmvdc, xlim = c(-2, 8), ylim = c(-2, 10)) ;

# CDF contour Plot

cdf=mvdc(copula = archmCopula(family = "gumbel", param = 1.4), margins = c("norm", "t"), paramMargins = list(list(mean = 0, sd = 1), list(df=3, ncp = 3)))

contour(cdf, pmvdc, xlim = c(-2, 20), ylim = c(-2, 20))

}

==============

# 3. Consider the monthly log returns of Boeing (BA), Abbott Labs (ABT),

# Mo- torola (MOT) and General Motors (GM) from January 1998 to December

# 2007. The log returns are in percentages.

# (a) Calculate the correlation co-efficient for MOT and GM.

# Generate 500 random number from gaussian copula taking parameter as

# this correlation coefficient. Draw contour plot of the empirical

# copula based on the generated sample and superimpose the empirical

# copula based on the data for MOT and GM.

corltn = function()

{

####### --------- Lybraries ------------#####

library("copula")

library("fCopulae")

data_set=read.table("m-ba4c9807.txt",header=TRUE);

BA=data_set[1];

ABT=data_set[2];

MOT=data_set[3];

GM=data_set[4];

MOTGM=data_set[3:4];

BA=t(BA);

ABT=t(ABT);

MOT=t(MOT);

GM=t(GM);

# 4 figures arranged in 2 rows and 2 columns

attach(mtcars)

par(mfrow=c(2,2))

rr=cor(t(MOT),t(GM), use = "everything", method = "pearson")

nc = normalCopula(-0.1231763)

r = rcopula(nc, 500)

# Probablity Empirical Copula:

empg=pempiricalCopula(r, N = 10)

contour(empg, pcopula, main="Probability Empirical contour plot of random gaussian data")

#par(new=TRUE)

empd=pempiricalCopula(MOTGM, N = 10)

contour(empd, pcopula, main="Probability Empirical contour plot of MOT and GM data")

# Density Empirical Copula:

empg=dempiricalCopula(r, N = 10)

contour(empg, dcopula, main="Density Empirical contour plot of random gaussian data")

#par(new=TRUE)

empd=dempiricalCopula(MOTGM, N = 10)

contour(empd, dcopula, main="Density Empirical contour plot of MOT and GM data")

}

===============================

****************** END ******************