如何用R语言基于死亡率表计算lx与dx？是否需用循环？

Hey there! Great question—you don’t have to use a for loop to compute dx and lx values. R’s vectorized operations can handle this efficiently (and are often faster for large datasets), but loops are totally a valid option if you prefer a more explicit, step-by-step workflow. Let’s walk through both methods using your data.

Step 1: Data Setup #

You’ve already loaded your mortality data with:

deathrate.table <- read.table("deathrate.txt", header = TRUE)

Let’s also define your initial life table value:

l1 <- 100000

Vectorized Approach (No Loop Needed) #

This method leverages R’s built-in functions to avoid manual iteration, which is the more idiomatic way to work in R:

1. First, calculate survival probabilities (1 minus the death rate) for each age:

deathrate.table$surv_prob <- 1 - deathrate.table$deathrate

2. Compute lx values using cumulative product of survival probabilities. Since each lx depends on the previous year’s survival, we can build the lx vector starting with l1 and multiplying by the cumulative product of prior survival rates:

# Start with l1, then append cumulative products scaled by l1
deathrate.table$lx <- c(l1, l1 * cumprod(deathrate.table$surv_prob[-nrow(deathrate.table)]))

3. Finally, calculate dx directly using your formula dx = lx * deathrate:

deathrate.table$dx <- deathrate.table$lx * deathrate.table$deathrate

For Loop Approach (Explicit Step-by-Step) #

If you want to see each iteration clearly (great for learning or debugging), a for loop works perfectly:

1. Initialize empty columns to store lx and dx:

deathrate.table$lx <- numeric(nrow(deathrate.table))
deathrate.table$dx <- numeric(nrow(deathrate.table))

2. Loop through each age to compute values step by step:

for (i in 1:nrow(deathrate.table)) {
  # Set initial lx for the first age
  if (i == 1) {
    deathrate.table$lx[i] <- l1
  } else {
    # Calculate current lx using previous lx minus previous dx
    deathrate.table$lx[i] <- deathrate.table$lx[i-1] - deathrate.table$dx[i-1]
  }
  # Calculate dx for the current age
  deathrate.table$dx[i] <- deathrate.table$lx[i] * deathrate.table$deathrate[i]
}

Key Takeaway #

Both methods will produce identical results. The vectorized approach is faster and more concise for large datasets, while the loop is easier to follow if you’re still getting comfortable with recurrence relations in R. Pick whichever fits your workflow best!

内容的提问来源于stack exchange，提问作者Dan Kingyens