CODEWITHSUNDEEP
r
Nick
Nick

Reputation: 53

Efficient recursive vector

I am trying to speed up a Monte Carlo function. I want to tighten this up but I can't figure out how. Any ideas?

initial=1400
ppinitial=0.07
assets2020=numeric(41)    
actual=round(seq(from=initial, to=0, by=(-initial/40)),0)*100000
assets2020[1]=1000000*1.02-actual[1]*ppinitial
for (i in 2:length(actual)){
  assets2020[i]=assets2020[i-1]*1.02-actual[i]*ppinitial}

Upvotes: 1

Views: 94

Answers (1)

agstudy
agstudy

Reputation: 121568

I think that you can use filter here:

assets2020 <- -ppinitial * filter(actual,1.02,'rec',init=-1000000/ppinitial)

demo:

y[i]=y[i-1]*1.02-x[i]*pp  ## use y ,x notation more handy than long variables

can be rewritten:

Y[i]= x[i]+f[1]*Y[i-1]  ## recursive filter expression

Where Y = -y/pp and f[1]=1.02

Or Y[1]=x[1]+f[1]*1000000 so init is equal to -1000000/ppinitial

PS : example for some:

ppinitial =100
filter_asset <- -ppinitial * filter(actual,1.02,'rec',init=-1000000/ppinitial)
sum(abs(filter_asset-assets2020))

[1] 0.001937866

Upvotes: 1

Related Questions