Skip to content

require help

12 messages · Ismail SEZEN, yadav neog, Berend Hasselman +3 more

yadav neog
# 
hello to all. I am working on macroeconomic data series of India, which in
a yearly basis. I am unable to convert my data frame into time series.
kindly help me.
also using zoo and xts packages. but they take only monthly observations.

'data.frame': 30 obs. of  4 variables:
 $ year: int  1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 ...
 $ cnsm: num  174 175 175 172 173 ...
 $ incm: num  53.4 53.7 53.5 53.2 53.3 ...
 $ wlth: num  60.3 60.5 60.2 60.1 60.7 ...
Eric Berger
# 
You did not provide the data frame so I will first create one and then use
it to create an xts

library(xts)
df <- data.frame( year=1980:2009, cnsm=sample(170:180,30,replace=TRUE),
                  incm=rnorm(30,53,1), wlth=rnorm(30,60,1))
dates <- as.Date(paste(df$year,"-01-01",sep=""))
myXts <- xts(df,order.by=dates)
On Fri, Sep 15, 2017 at 12:38 PM, yadav neog <yadavneog at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      Ismail SEZEN
    

    

      
      #
    

  


  

          

      
      
      
    

        
Do you really need to convert your data to time series/xts/zoo? I don?t know you try what kind of an analysis but perhaps you don?t have to.

            

      
      
      
    

        
If you really have to convert to xts/zoo, why don?t yo set each year to first day of January and use it as is? For instance,

index, cnsm, incm, wlth
1980-01-01, 174, 53.4, 60.3
1981-01-01, 175, 53.7, 60.5
1982-01-01, 175, 53.5, 60.2
?..

            

      
      
      
    

  
  


  


      

    

    

      
      

        


  

        

      yadav neog
    

    

      
      #
    

  


  

      
thanks, eric../ actually I have the data which have not specify the months.
therefore i bound to declare is in yearly data. i also attached a sample
data set that may be helpful for you to providing suggestions. thank you

        
On Fri, Sep 15, 2017 at 5:23 PM, Ismail SEZEN <sezenismail at gmail.com> wrote:

        

            

      
      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Berend Hasselman
    

    

      
      #
    

  


  

          

      
      
      
    

        
It shouldn't be difficult.
Example:

tsdata <- data.frame(year=c(2000,2002,2003), x=c(1,2,3),y=c(10,11,12))
xy.ts <- as.ts(tsdata)

library(zoo)

as.zoo(xy.ts)


Berend Hasselman

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Berend Hasselman
    

    

      
      #
    

  


  

          

      
      
      
    

        
Ignore my suggestion.  Doesn't do what you need.

Berend

  
  


  


      

    

    

      
      

        


  

        

      Berend Hasselman
    

    

      
      #
    

  


  

          

      
      
      
    

        
Second try to do what you would like (I hope and think)
Using Eric's sample data

<code>
zdf <- data.frame(year=2001:2010, cnsm=sample(170:180,10,replace=TRUE),
                 incm=rnorm(10,53,1), wlth=rnorm(10,60,1))
zdf

# R ts
zts <- ts(zdf[,-1], start=zdf[1,"year"])
zts

# turn data into a zoo timeseries and an xts timeseries

library(zoo)
z.zoo <- as.zoo(zts)
z.zoo

library(xts)
z.xts <- as.xts(zts)
z.xts
</code>

Berend Hasselman

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Eric Berger
    

    

      
      #
    

  


  

      
You can just use the same code that I provided before but now use your
dataset. Like this

df <- read.csv(file="data2.csv",header=TRUE)
dates <- as.Date(paste(df$year,"-01-01",sep=""))
myXts <- xts(df,order.by=dates)
head(myXts)

#The last command "head(myXts)" shows you the first few rows of the xts
object
           year     cnsm    incm    wlth
1980-01-01 1980 173.6527 53.3635 60.3013
1981-01-01 1981 175.4613 53.6929 60.4980
1982-01-01 1982 174.5724 53.4890 60.2358
1983-01-01 1983 171.5070 53.2223 60.1047
1984-01-01 1984 173.3462 53.2851 60.6946
1985-01-01 1985 171.7075 53.1596 60.7598

        
On Sat, Sep 16, 2017 at 9:55 AM, Berend Hasselman <bhh at xs4all.nl> wrote:

        

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      yadav neog
    

    

      
      #
    

  


  

      
oky.. thank you very much to all of you

        
On Sat, Sep 16, 2017 at 2:06 PM, Eric Berger <ericjberger at gmail.com> wrote:

        

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

  
    

      
      
    

  


  


      

    

    
5 days later

    

      
      

        


  

        

      Gabor Grothendieck
    

    

      
      #
    

  


  

      
Assuming the input data.frame, DF, is of the form  shown reproducibly
in the Note below, to convert the series to zoo or ts:

library(zoo)

# convert to zoo
z <- read.zoo(DF)

# convert to ts
as.ts(z) #


Note:

DF <- structure(list(year = c(1980, 1981, 1982, 1983, 1984), cnsm = c(174,
175, 175, 172, 173), incm = c(53.4, 53.7, 53.5, 53.2, 53.3),
    with = c(60.3, 60.5, 60.2, 60.1, 60.7)), .Names = c("year",
"cnsm", "incm", "with"), row.names = c(NA, -5L), class = "data.frame")

        
On Sat, Sep 16, 2017 at 8:10 AM, yadav neog <yadavneog at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      yadav neog
    

    

      
      #
    

  


  

      
thankx to everyone for your valuable suggestions. one query regarding the
GARCH model.
I have applied the GARCH model for the same data that I send you all . and
my results coming like

Error in .sgarchfit(spec = spec, data = data, out.sample = out.sample,  :

ugarchfit-->error: function requires at least 100 data
 points to run

can you suggest something on it.

On Fri, Sep 22, 2017 at 6:02 AM, Gabor Grothendieck <ggrothendieck at gmail.com

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Joshua Ulrich
    

    

      
      #
    

  


  

      
On Fri, Sep 22, 2017 at 7:28 AM, yadav neog <yadavneog at gmail.com> wrote:

            

      
      
      
    

        
The error is protecting you from the unreliable results that you would
get by running the model on too few observations.  You either need
more data, or a different algorithm (not GARCH).