创建稀疏距离矩阵

Question

我有以下向量：

a <- 1:10
b <- c(100, 300, 400, 120, 131,
       450, 400, 432, 111, 123)
rep_ab <- unlist(lapply(a, function(x) rep(x, times = b[x])))
length(rep_ab) == sum(b)
# TRUE

I would like to create a distance matrix using the vector `rep_ab`. Usually, I would use the following code:

mat <- outer(rep_ab, rep_ab, FUN = "-")

但是由于这个矩阵太大，我想使用稀疏矩阵创建这个距离矩阵。我有办法做到这一点吗？

提前致谢。

Answer 1

library(Matrix)

a <- 1:10
b <- c(100, 300, 400, 120, 131, 450, 400, 432, 111, 123)
rep_ab <- unlist(lapply(a, function(x) rep(x, times = b[x])))

# Create a sparse matrix with 0 values
sparse_matrix <- Matrix(0, nrow = length(rep_ab), ncol = length(rep_ab), sparse = TRUE)

# Fill the sparse matrix with distance values
for (i in 1:length(rep_ab)) {
  for (j in 1:length(rep_ab)) {
    sparse_matrix[i, j] <- abs(rep_ab[i] - rep_ab[j])
  }
}

# Convert to a sparse distance matrix
sparse_distance_matrix <- as(sparse_matrix, "CsparseMatrix")
print(sparse_distance_matrix)