Weighted Mean in R-CodePudding

I have a time-series data with different lengths - a,b,c,d,e

 #printing a
 Time Series:
  Start = 1
  End = 42
  Frequency = 1
  [1] 60 
 
#printing b
Time Series:
  Start = 1
  End = 42
  Frequency = 1
  [1] 50 70  
 
#printing c
Time Series:
  Start = 1
  End = 42
  Frequency = 1
  [1] 40 70 100

 #and so on

I am trying to get the mean of elements in all lists: #since there are 5 values available for 1st element

mean1 <- a[1] b[1] c[1] d[1] e[1] / 5 

 #since there are 4 values available for 2nd element

mean2<-  b[2] c[2] d[2] e[2] / 4

#next divide by 3 and 2...1

mean3<- c[3] d[3] e[3] / 3 and so on...

I need the mean of these values so that I can make a weighted mean for each element for further processing. Can anyone give suggestion on what to do to obtain the weighted mean from this data??

CodePudding user response：

We get the objects in a list, then get the mean from the list and do the weighted calculation

lst1 <- lapply(mget(letters[1:5]), unlist)
mx <- max(lengths(lst1))
lst2 <- lapply(lst1, `length<-`, mx)
(1/mx) * (rowSums(mapply(`*`, lst2,  
     rowMeans(simplify2array(lst2), na.rm = TRUE)), na.rm = TRUE))

CodePudding user response：

you could do:

l <- c(a,b,c,d,e)

tapply(unlist(l), sequence(lengths(l)), mean)
        1         2         3         4         5 
 6923.783 -2462.537 16402.663    62.005   432.800