Reg: Frequency in declaring time series data
Hi Upananda, A few comments: 1. As you know, CRAN has thousands of packages. One of the ways to learn about the packages you might care about is to use the CRAN views. A 'view' is an attempt to provide some information on a certain subset of the packages related to a particular area. See a list of the views at: https://cran.r-project.org/web/views/ 2. In your case, you might be interested in the 'Time Series' view: https://cran.r-project.org/web/views/TimeSeries.html 3. On that page, the subsection 'Time Series Classes' mentions two packages, 'tsbox' and 'timetk' which provide tools to convert between different representations of time series used in different packages. I have not used either of these packages, but perhaps they could be of use to you. 4. For financial time series, it is often useful to have the exact dates. Rui mentioned the zoo package. I often use the xts package. xts = eXtensible Time Series. The xts package is an extension of the zoo package. 5. If your data is coming from market prices, and you have daily data corresponding to market prices, then presumably you don't have weekend or holiday prices (when the markets are closed). This leads to about 250 or 252 trading days per year (52 weeks X 5 days/wk = 260, minus 8-10 holidays). It is standard in financial calculations to annualize based on 250 or 252. So, you might want to specify a frequency of 250 or 252 in your time series when creating a ts object. I hope that helps, Eric
On Tue, Jan 17, 2023 at 6:29 PM Upananda Pani <upananda.pani at gmail.com> wrote:
Hi Rui, Thank you so much for your help. As I have to fit a Markov Switching Model using MSwM package. May I know whether i can convert from zoo object to a time series object. As I have to use several packages which uses ts so I am not able to decide how to do it. Grateful to you for your help and support. With sincere regards, Upananda On Tue, 17 Jan, 2023, 01:40 Rui Barradas, <ruipbarradas at sapo.pt> wrote:
?s 16:39 de 16/01/2023, Upananda Pani escreveu:
Dear All, I have a time series daily data with date are stored ( %dd-%mm-%yy
format )
from 22-01-20 to 03-08-21. In total I have 560 observations. I am using
the
following command to declare as a time series object. Here the the data
set
is 7 days a week.
oil <- read_xlsx("crudefinal.xlsx")
pricet=ts(oil$price, start = c(2020, 22), freq = 365)
roilt=ts(diff(log(oil$price))*100,start=c(2020,22), freq=365)
Shall I have to declare the dates here? I want to know also if it is a 5
day trading a week, how to declare the frequency.
Looking forward to your reply
Regards,
Upananda Pani
Looking forward to your suggestions.
[[alternative HTML version deleted]]
______________________________________________ R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. Hello, Package zoo is the best way of having dates in a time series. The package comes with several vignettes that can get you started quickly. See the difference between classes ts and zoo below. # make up some test data oil <- data.frame(price = cumsum(rnorm(560))) oil$date <- seq(as.Date("2020-01-22"), by = "1 day", length.out = 560) # base R pricet <- ts(oil$price, start = c(2020, 22), freq = 365) time(pricet) index(pricet) plot(pricet) #--- library(zoo) library(ggplot2) pricez <- zoo(oil$price, order.by = oil$date) time(pricez) index(pricez) autoplot(pricez) vignette(package = "zoo") Hope this helps, Rui Barradas
[[alternative HTML version deleted]]
______________________________________________ R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.