问题
I want to calculate normalised ratios and simple ratios in all possible combinations in R. This is the sample dataset
df = structure(list(var_1 = c(0.035, 0.047, 0.004, 0.011, 0.01, 0.01,
0.024), var_2 = c(0.034, 0.047, 0.004, 0.012, 0.01, 0.011, 0.025
), var_3 = c(0.034, 0.047, 0.006, 0.013, 0.011, 0.013, 0.026),
var_4 = c(0.034, 0.046, 0.008, 0.016, 0.014, 0.015, 0.028
), var_5 = c(0.034, 0.046, 0.009, 0.017, 0.015, 0.016, 0.029
)), class = "data.frame", row.names = c(NA, -7L))
I could able to calculate simple ratios in all possible combinations after taking help from this.
do.call("cbind", lapply(seq_along(df), function(y) apply(df, 2, function(x) df[[y]]/x)))
But I am unable to calculate normalised ratios i.e. (xj - xi)/(xj + xi) and how to name each calculated ratios properly?
回答1:
Perhaps, you can try nested lapply to get all the combinations :
cols <- 1:ncol(df)
mat <- do.call(cbind, lapply(cols, function(xj)
sapply(cols, function(xi) (df[, xj] - df[, xi])/(df[, xj] + df[, xi]))))
To assign column names, we can use outer
colnames(mat) <- outer(names(df), names(df), paste0)
Thinking about it I think we can directly manipulate this using column indexes.
cols <- 1:ncol(df)
temp <- expand.grid(cols, cols)
new_data <- (df[,temp[,2]] - df[,temp[,1]])/(df[,temp[,2]] + df[,temp[,1]])
回答2:
We could do this more easily with outer alone
f1 <- function(i, j) (df[, i] - df[, j])/(df[, i] + df[, j])
out <- outer(seq_along(df), seq_along(df), FUN = f1)
colnames(out) <- outer(names(df), names(df), paste0)
来源：https://stackoverflow.com/questions/62915303/how-to-calculate-ratios-and-normalized-ratios-in-all-possible-combinations-in-r