Skip to content

Help with interpolation

2 messages · beanbandit, Jessica Streicher

Original
beanbandit
# 
hi guys

I need to interpolate values for the zero coupon yield curve. Following data
is given

                                                                                                                                                                                                                                                                
 
    date            days      rate

1996 01 02      15    5.74590
1996 01 02      50    5.67332
1996 01 02      78    5.60888
1996 01 02     169    5.47376
1996 01 02     260    5.35267
1996 01 02     351    5.27619

1996 01 03      14    5.74740
1996 01 03      49    5.67226
1996 01 03      77    5.60371
1996 01 03     168    5.47058
1996 01 03     259    5.34662
1996 01 03     350    5.26630
 
For every day i have to interpolate 10 values, for example for maturities of
30,60 or 90 days. I have interpolate data for a one year period, 10
interpolation values a day, so that equals 3600 values. 

what's the easiest way to implement this in R?

please hlep!




--
View this message in context: http://r.789695.n4.nabble.com/Help-with-interpolation-tp4655843.html
Sent from the R help mailing list archive at Nabble.com.
4 days later
Jessica Streicher
# 
Next time please provide sample data in a form we can easily read in (look at ?dput for example)

If i understand this right:

yourData<-read.table(header=T,text="
date			days  rate
1996_01_02      15    5.74590
1996_01_02      50    5.67332
1996_01_02      78    5.60888
1996_01_02     169    5.47376
1996_01_02     260    5.35267
1996_01_02     351    5.27619

1996_01_03      14    5.74740
1996_01_03      49    5.67226
1996_01_03      77    5.60371
1996_01_03     168    5.47058
1996_01_03     259    5.34662
1996_01_03     350    5.26630
")

results<-sapply(unique(yourData$date),function(thisDate){
			subSet <- yourData[yourData$date==thisDate,]
			appr<-approx(subSet$days,subSet$rate,xout=seq(0,360, by=30))
			rates<-appr$y
			names(rates)<-appr$x
			rates
		})
colnames(results)<-unique(yourData$date)

This gives 13 results per date though, and it can't interpolate the first and last value. If you need those values that are not in-between, try spline instead of approx (you never specified how you wanted to interpolate).
On 17.01.2013, at 15:50, beanbandit wrote: