R语言中按0.8阈值返回对应族群列名的续问需求

Hey there! Let's adjust your code to meet the new requirement—only return an ethnicity if its proportion is over 0.8, otherwise fall back to "No Match". Here's how to do it step by step:

Solution Overview #

The core logic we need to implement is:

• For each individual, find the maximum proportion across all ethnicity columns
• Check if this maximum value is ≥ 0.8
  • If yes: return the corresponding ethnicity name
  • If no: return "No Match"

Base R Approach (No Extra Packages Needed) #

First, define the columns that hold ethnicity proportion data:

# List of columns with ethnicity percentage values
ethnic_cols <- c("white", "black", "hispanic", "asian", "other")

Then, calculate the max percentage per row, map it to the corresponding ethnicity, and apply our condition:

# Get the highest percentage for each row
max_pct <- apply(ethnicity[ethnic_cols], 1, max)

# Find the ethnicity associated with that highest percentage (matches your original logic)
top_ethnic <- ethnic[max.col(ethnicity[ethnic_cols], 'first')]

# Update the ethnicity column with our new rule
ethnicity$ethnicity <- ifelse(max_pct >= 0.8, top_ethnic, "No Match")

dplyr/Tidyverse Approach (More Readable) #

If you prefer tidyverse syntax, this version makes the row-by-row logic more explicit:

library(dplyr)

ethnicity <- ethnicity %>%
  rowwise() %>%
  mutate(
    # Calculate the maximum percentage across all ethnicity columns
    max_pct = max(c_across(all_of(ethnic_cols))),
    # Identify which ethnicity has that maximum percentage
    top_ethnic = ethnic[which.max(c_across(all_of(ethnic_cols)))],
    # Apply our condition to set the final ethnicity value
    ethnicity = ifelse(max_pct >= 0.8, top_ethnic, "No Match")
  ) %>%
  ungroup()

Verify with Your Sample Data #

Let's check how this works with your provided dataset:

• Row 1998: hispanic is 0.8306 (≥0.8) → returns "hispanic"
• Row 3431: Max percentage is 0.5963 (white) <0.8 → returns "No Match"
• Row 9524: asian is 0.8649 (≥0.8) → returns "asian"

This aligns perfectly with the behavior you described!

内容的提问来源于stack exchange，提问作者user12801590