RStudio使用kohonen包构建SOM模型时遇对象未找到等错误求助

Let’s break down your issues one by one—these are super common pitfalls when starting out with self-organizing maps in R, so you’re not alone!

The Root Cause: You’re Relying on R’s Built-in data Function #

Your first error is the domino that causes all the others:

Error in data[, c(2, 4, 5, 8)] : object of type 'closure' is not subsettable

data is a built-in R function (used to load pre-packaged datasets), not your custom dataset. You forgot to load or define your own data, so R tried to treat the data() function like a dataset—which doesn’t work.

Step-by-Step Fixes #

Let’s rewrite your code with corrections that address every error:

1. Load Your Actual Dataset First
   Replace the generic data with your real dataset. For example, if your data is in a CSV:

   # Load your data (update the file path to match your file)
   my_custom_data <- read.csv("your_data_file.csv")
   

   If your data is already in your R environment, use its actual name instead of my_custom_data.

2. Create Your Training Subset
   Now use your named dataset to build data_train:

   require(kohonen)
   
   # Use your actual dataset name here
   data_train <- my_custom_data[, c(2,4,5,8)]
   

3. Generate the Scaled Matrix
   With data_train properly created, this line will run without missing object errors:

   data_train_matrix <- as.matrix(scale(data_train))
   

4. Train the SOM and Visualize Results
   The rest of your code will work now that all prerequisite objects exist:

   som_grid <- somgrid(xdim = 20, ydim=20, topo="hexagonal")
   som_model <- som(data_train_matrix, grid=som_grid, rlen=500, alpha=c(0.05,0.01), keep.data = TRUE )
   plot(som_model, type="changes")
   

Why the Tutorial Code Worked #

Tutorials almost always assume you’ve already loaded a dataset named data (or they explicitly define it in the example). You skipped that critical step of loading your own data and accidentally used R’s built-in data function instead.

内容的提问来源于stack exchange，提问作者caroline