我正在尝试计算3列的加权平均值,其中根据每行缺失值的数量来确定权重。
一个可重复的例子:
# Some simulated data
N <- 50
df <- data.table(int_1 = runif(N,1000,5000), int_2 = runif(N,1000,5000), int_3 = runif(N,1000,5000))
df[-1] <- lapply(df[-1], function(x) { x[sample(c(1:N), floor(N/10))] <- NA ; x })
# Function to calculate weighted average
# The weights are flexible and are input by user
a = 5
b = 3
c = 2
i = 10
wa_func <- function(x,y,z){
if(!(is.na(x) & is.na(y) & is.na(z))){
wt_avg <- (a/i)* x + (b/i) * y + (c/i) * z
} else if(!is.na(x) & !is.na(y) & is.na(z)){
wt_avg <- (a/(i-c))* x + (b/(i-c)) * y
} else if(!is.na(x) & is.na(y) & is.na(z)){
wt_avg <- a/(i-(b+c))* x
}
return(wt_avg)
}
df[, weighted_avg_int := mapply(wa_func,int_1,int_2,int_3)]
但是该函数输出NA用于连续的任何缺失值。我在这里错过了什么?
提前致谢。
您需要在函数中更改第一个if
的条件:
wa_func <- function(x, y, z) {
if (!(is.na(x) | is.na(y) | is.na(z))) {
wt_avg <- (a / i) * x + (b / i) * y + (c / i) * z
} else if (!is.na(x) & !is.na(y) & is.na(z)) {
wt_avg <- (a / (i - c)) * x + (b / (i - c)) * y
} else if (!is.na(x) & is.na(y) & is.na(z)) {
wt_avg <- a / (i - (b + c)) * x
}
return(wt_avg)
}
你可以通过用qazxsw poi包装你的函数来改进函数,所以你不需要qazxsw poi:
mapply