[R-SIG-Finance] Fwd: runMult instead of runSum
G See
gsee000 at gmail.com
Sun Oct 30 01:14:23 CEST 2011
Hi Martin,
Attached is a script I made for you to demonstrate a few of the many
possible approaches you could take. Below, I pasted the output from the
script.
HTH,
Garrett
R> # There are lots of ways to do this. First, I'll do something similar to
R> # what you have. i.e. use the "monthlyReturn" function and do each
stock separately
R>
R> library(quantmod)
R> s <- c("DD", "DIS", "PFE")
R> getSymbols(s)
[1] "DD" "DIS" "PFE"
R> ### 1
R>
R> # use leading=FALSE so that monthlyReturn won't replace the first
observation
R> # The first row will be NA, so use -1 subset to remove it
R> DD.m1 <- monthlyReturn(Ad(DD), leading=FALSE)[-1,]
R> # apply a rolling function to convert 1 month returns to 2 and 3 month
returns
R> DD.m2 <- rollapplyr(DD.m1, width=2, FUN=function(x) prod(1 + x) - 1)
R> DD.m3 <- rollapplyr(DD.m1, width=3, FUN=function(x) prod(1 + x) - 1)
R> # then you'd do this for each symbol
R>
R> ### 2
R>
R> # Or, you can use log returns so that you can use the runSum function,
then "unlog" them at the end
R> DD.m <- monthlyReturn(Ad(DD), type='log', leading=FALSE)[-1,]
R> DD.m1 <- exp(runSum(DD.m, n=1)) - 1
R> DD.m2 <- exp(runSum(DD.m, n=2)) - 1
R> DD.m3 <- exp(runSum(DD.m, n=3)) - 1
R>
R> ### 3
R>
R> # Alternatively, you could convert the prices to monthly first, and then
calculate the returns
R> mDD <- to.monthly(DD)
R> # For univariate series, you can use Delt or ROC
R> DD.m1 <- Delt(Ad(mDD))
R> DD.m2 <- ROC(Ad(mDD), n=2, type='discrete')
R> DD.m3 <- Delt(Ad(mDD), k=3)
R> # then you'd do this for each symbol
R>
R> ### 4
R>
R> # Since you're only dealing with the Adjusted price of each Symbol, it's
probably better
R> # to merge all the relevant prices into a data.frame
R> p <- cbind(Ad(DD), Ad(DIS), Ad(PFE))
R> # convert to monthly prices
R> mp <- to.monthly(p, OHLC=FALSE)
R> # calculate 1 month, 2 month, and 3 month returns with ROC
R> ret1 <- ROC(mp, n=1, type='discrete')
R> ret2 <- ROC(mp, n=2, type='discrete')
R> ret3 <- ROC(mp, n=3, type='discrete')
R> head(ret1)
DD.Adjusted DIS.Adjusted PFE.Adjusted
Jan 2007 NA NA NA
Feb 2007 0.031112200 -0.025954654 -0.03851709
Mar 2007 -0.025659301 0.005206738 0.01201803
Apr 2007 -0.005120702 0.015843998 0.04750124
May 2007 0.071813725 0.013197361 0.05007085
Jun 2007 -0.028355820 -0.036708111 -0.06972560
R> head(ret2)
DD.Adjusted DIS.Adjusted PFE.Adjusted
Jan 2007 NA NA NA
Feb 2007 NA NA NA
Mar 2007 0.004654581 -0.02088305 -0.02696196
Apr 2007 -0.030648610 0.02113323 0.06009014
May 2007 0.066325287 0.02925046 0.09995052
Jun 2007 0.041421569 -0.02399520 -0.02314596
R> # for many different return periods, you can use lapply (or a for loop
if you prefer)
R> rets <- lapply(1:5, function(n) ROC(mp, n=n, type='discrete'))
R> head(rets[[1]]) # 1 month returns
DD.Adjusted DIS.Adjusted PFE.Adjusted
Jan 2007 NA NA NA
Feb 2007 0.031112200 -0.025954654 -0.03851709
Mar 2007 -0.025659301 0.005206738 0.01201803
Apr 2007 -0.005120702 0.015843998 0.04750124
May 2007 0.071813725 0.013197361 0.05007085
Jun 2007 -0.028355820 -0.036708111 -0.06972560
R> head(rets[[2]]) # 2 month returns
DD.Adjusted DIS.Adjusted PFE.Adjusted
Jan 2007 NA NA NA
Feb 2007 NA NA NA
Mar 2007 0.004654581 -0.02088305 -0.02696196
Apr 2007 -0.030648610 0.02113323 0.06009014
May 2007 0.066325287 0.02925046 0.09995052
Jun 2007 0.041421569 -0.02399520 -0.02314596
R> # etc.
R>
R> # If you have a lot of symbols, you may be interested in the "PF"
function from
R> # my "qmao" package (which is on R-Forge) which will do the "cbind"ing
for you.
R> # install.packages("qmao", repos='http://r-forge.r-project.org')
R> # Just to have a little fun, here is how you could quickly calculate 6
month returns for all Dow components
R>
R> library(qmao)
R> library(XML)
R> #First get the symbols
R> DJIsymbols <- function() {
R+ djicomp <-
readHTMLTable('http://finance.yahoo.com/q/cp?s=^DJI+Components',
as.data.frame=FALSE)
R+ as.data.frame(djicomp[[10]], stringsAsFactors=FALSE)[,1]
R+ }
R> syms <- DJIsymbols()
R> getSymbols(syms)
<snip>
[1] "AA" "AXP" "BA" "BAC" "CAT" "CSCO" "CVX" "DD"
[9] "DIS" "GE" "HD" "HPQ" "IBM" "INTC" "JNJ" "JPM"
[17] "KFT" "KO" "MCD" "MMM" "MRK" "MSFT" "PFE" "PG"
[25] "T" "TRV" "UTX" "VZ" "WMT" "XOM"
R> P <- PF(syms, prefer='Adjusted', silent=TRUE)
R> mP <- to.monthly(P, OHLC=FALSE)
R> # 6 month simple return of each Dow component
R> tail(ROC(mP, n=6, type='discrete'))
AA AXP BA BAC
May 2011 0.28407351 0.2035731077 0.23722861 0.0751604
Jun 2011 0.03464052 0.2186982249 0.14537514 -0.1766917
Jul 2011 -0.10800971 0.1623512946 0.02564103 -0.2914536
Aug 2011 -0.23718713 0.1495587552 -0.06029515 -0.4266667
Sep 2011 -0.45594088 0.0008954556 -0.17166324 -0.5398496
Oct 2011 -0.31659776 0.0698151951 -0.13522771 -0.4000000
CAT CSCO CVX DD
May 2011 0.2616811 -0.12031662 0.31547017 0.15129714
Jun 2011 0.1467261 -0.22550000 0.14423508 0.10002054
Jul 2011 0.0269677 -0.23912004 0.11244071 0.02991712
Aug 2011 -0.1084622 -0.14931880 -0.03230860 -0.10561423
Sep 2011 -0.3312995 -0.09283196 -0.12502363 -0.26077307
Oct 2011 -0.1525943 0.06789413 0.01763505 -0.11636233
DIS GE HD HPQ
May 2011 0.152866242 0.25913838 0.20906631 -0.1050471
Jun 2011 0.040789123 0.04708520 0.04779948 -0.1306122
Jul 2011 -0.006431695 -0.09720102 -0.03640912 -0.2262774
Aug 2011 -0.221307727 -0.21362530 -0.08971912 -0.4002780
Sep 2011 -0.300069622 -0.22741117 -0.09969871 -0.4471805
Oct 2011 -0.159860789 -0.14136386 -0.01311475 -0.3018491
IBM INTC JNJ JPM
May 2011 0.20426720 0.08098933 0.11277945 0.16361650
Jun 2011 0.17881297 0.07125427 0.09463722 -0.02866089
Jul 2011 0.13197874 0.05738881 0.10324686 -0.08946412
Aug 2011 0.07135735 -0.04597156 0.08994534 -0.18622951
Sep 2011 0.08184855 0.07614725 0.09395397 -0.33894178
Oct 2011 0.10864679 0.09802198 0.01579436 -0.18394128
KFT KO MCD MMM
May 2011 0.17773973 0.06532988 0.05772496 0.137328642
Jun 2011 0.13811075 0.03789408 0.11562542 0.112147326
Jul 2011 0.14472801 0.09783845 0.19239311 0.003012746
Aug 2011 0.11960026 0.11808885 0.21160547 -0.088932806
Sep 2011 0.08884565 0.03271171 0.17062117 -0.222462905
Oct 2011 0.07175295 0.03622971 0.20826318 -0.156162100
MRK MSFT PFE
May 2011 0.0901365706 0.0024203308 0.342405063
Jun 2011 0.0014359563 -0.0565899963 0.200353565
Jul 2011 0.0521060842 0.0003673769 0.078142695
Aug 2011 0.0394026057 0.0137195122 0.006896552
Sep 2011 0.0133250697 -0.0071798963 -0.111111111
Oct 2011 -0.0008537279 0.0543180930 -0.035053554
PG T TRV UTX
May 2011 0.114454776 0.16844512 0.16457143 0.17840249
Jun 2011 0.003691221 0.09985580 0.06161746 0.13621091
Jul 2011 -0.009902597 0.09332322 -0.00744507 0.02965465
Aug 2011 0.026820546 0.03200883 -0.14726436 -0.10032715
Sep 2011 0.042588588 -0.04195089 -0.16814612 -0.15867512
Oct 2011 0.013782302 -0.01620906 -0.04572437 -0.10158192
VZ WMT XOM
May 2011 0.183733948 0.040425128 0.2131675201
Jun 2011 0.067766647 -0.001135933 0.1252609603
Jul 2011 0.016647196 -0.047159505 -0.0001261352
Aug 2011 0.005071851 0.044784915 -0.1239199905
Sep 2011 -0.020248380 0.010907674 -0.1260979425
Oct 2011 0.023110386 0.053844735 -0.0624784260
On Sat, Oct 29, 2011 at 3:42 PM, Martin Bauer <Bauermartin at gmx.at> wrote:
> -------- Original-Nachricht --------
> Datum: Fri, 28 Oct 2011 18:13:14 +0200
> Von: "Martina Bauer" <Bauermartin at gmx.at>
> An: r-sig-finance at r-project.or
> Betreff: runMult instead of runSum
>
> Hello,
>
> Maybe my approach is wrong from the beginning but I would like to
> calculate from monthly data - one months and two months and lastly three
> months return
> I know that I have to multiply the monthly performance instead to sum up -
> but I dont know how to do in R. Finally I should also rank the result
> according to n period performance (no clue how to archive this in R)
>
>
> here my code
>
> require(quantmod)
> stock=c("DD","DIS","PFE")
> DD.sp=Ad(DD)
> dd.m=monthlyReturn(DD.sp)
> dd1=runSum(dd.m,n=1)
> dd2=runSum(dd.m,n=2)
> dd3=runSum(dd.m,n=3)
>
> DIS.sp=Ad(DIS)
> dis.m=monthlyReturn(DIS.sp)
> dis1=runSum(DIS.m,n=1)
> dis2=runSum(DIS.m,n=2)
> dis3=runSum(DIS.m,n=3)
>
> PFE.sp=Ad(PFE)
> pfe.m=monthlyReturn(PFE.sp)
> pfe1=runSum(pfe.m,n=1)
> pfe2=runSum(pfe.m,n=2)
> pfe3=runSum(pfe.m,n=3)
>
>
> I know that my R code is very poor any idea how to improve ?
>
> THanks in advance
> Best Regards
>
> martin
>
> --
>
>
>
>
> --
>
> _______________________________________________
> R-SIG-Finance at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only. If you want to post, subscribe first.
> -- Also note that this is not the r-help list where general R questions
> should go.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://stat.ethz.ch/pipermail/r-sig-finance/attachments/20111029/82253abe/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: MartinReturns.R
Type: text/x-r-source
Size: 2632 bytes
Desc: not available
URL: <https://stat.ethz.ch/pipermail/r-sig-finance/attachments/20111029/82253abe/attachment.bin>
More information about the R-SIG-Finance
mailing list