labeled inverse cell frequency: relative frequency
idf_rf(expr, features = NULL, label, multi = TRUE, thres = 0)a matrix, features in row and cells in column
vector, feature names or indexes to compute
vector, group label of each cell
logical, if to compute based on binary (FALSE) or multi-class (TRUE)
numeric, cell only counts when expr > threshold, default 0
a matrix of inverse cell frequency score
$$\mathbf{IDF_{i,j}} = log(1+\frac{\frac{n_{i,j\in D}}{n_{j\in D}}}{max(\frac{n_{i,j\in \hat D}}{n_{j\in \hat D}})+ e^{-8}})$$ where \(n_{i,j\in D}\) is the number of cells containing feature \(i\) in class \(D\), \(n_{j\in D}\) is the total number of cells in class \(D\), \(\hat D\) is the class except \(D\).
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf_rf(data, label = sample(c("A", "B"), 10, replace = TRUE))
#> B B A B A B A
#> 1 0.5596158 0.5596158 0.8472979 0.5596158 0.8472979 0.5596158 0.8472979
#> 2 0.9162907 0.9162907 0.5108256 0.9162907 0.5108256 0.9162907 0.5108256
#> 3 0.6931472 0.6931472 0.6931472 0.6931472 0.6931472 0.6931472 0.6931472
#> 4 0.6418539 0.6418539 0.7472144 0.6418539 0.7472144 0.6418539 0.7472144
#> 5 0.7884574 0.7884574 0.6061358 0.7884574 0.6061358 0.7884574 0.6061358
#> 6 0.7884574 0.7884574 0.6061358 0.7884574 0.6061358 0.7884574 0.6061358
#> 7 0.6418539 0.6418539 0.7472144 0.6418539 0.7472144 0.6418539 0.7472144
#> 8 0.6418539 0.6418539 0.7472144 0.6418539 0.7472144 0.6418539 0.7472144
#> 9 0.6418539 0.6418539 0.7472144 0.6418539 0.7472144 0.6418539 0.7472144
#> 10 0.5596158 0.5596158 0.8472979 0.5596158 0.8472979 0.5596158 0.8472979
#> A A A
#> 1 0.8472979 0.8472979 0.8472979
#> 2 0.5108256 0.5108256 0.5108256
#> 3 0.6931472 0.6931472 0.6931472
#> 4 0.7472144 0.7472144 0.7472144
#> 5 0.6061358 0.6061358 0.6061358
#> 6 0.6061358 0.6061358 0.6061358
#> 7 0.7472144 0.7472144 0.7472144
#> 8 0.7472144 0.7472144 0.7472144
#> 9 0.7472144 0.7472144 0.7472144
#> 10 0.8472979 0.8472979 0.8472979