如何在R语言中非随机拆分训练集与测试集（取前80%为训练集）

Hey there! Let's work through fixing your code to split your data into a training set (first 80% of rows) and test set (remaining 20%) correctly.

First, let's spot the small issues in your original code:

• Typos: You defined rows = nrow(data) but then used 0.80*row (missing an 's' at the end of row)
• Non-integer index: Calculating 0.8 * rows might give a decimal number, but R requires integer values for row indexing—we need to convert that to a whole number
• Incorrect subset syntax: data[1:index] selects columns instead of rows. To grab rows, you need to include a comma to indicate "all columns": data[1:index, ]

Corrected Base R Code #

Here's the fixed version that addresses all these issues:

# Get total number of rows in your dataset
total_rows <- nrow(data)

# Calculate the last row index for the training set (use floor() to get an integer)
train_last_row <- floor(0.8 * total_rows)

# Split the data
train_data <- data[1:train_last_row, ]  # First 80% of rows, all columns
test_data <- data[(train_last_row + 1):total_rows, ]  # Remaining rows, all columns

Alternative with Tidyverse (dplyr) #

If you prefer using the tidyverse ecosystem, the slice() function makes this even more readable:

library(dplyr)

# Calculate the split index (same as before)
train_last_row <- floor(0.8 * nrow(data))

# Split using slice()
train_data <- data %>% slice(1:train_last_row)
test_data <- data %>% slice((train_last_row + 1):n())  # n() gets the total number of rows

Just remember to replace data with the actual name of your data frame!

内容的提问来源于stack exchange，提问作者user15051990