Score the uLSIF density-ratio LRT discriminator — score_dre

Scores each row of X independently with the frozen model learned by fit_dre_ulsif. For one specimen the abundances are mapped to the self-contained per-sample rCLR representation $z$, and the score is the log estimated density ratio $$S(z) = \log\max\!\Bigl(\sum_l \alpha_l \exp\bigl(-\lVert z - c_l\rVert^2 / (2\sigma^2)\bigr),\, \varepsilon\Bigr),$$ with larger values more case-like.

At scoring time the present feature set is intersect(model$feature_universe, colnames(X)). The kernel centres, bandwidth and coefficients are re-derived over exactly that present set from the frozen raw anchors / centres – using only frozen training data, never a scored-batch statistic – which keeps partial-overlap scores consistent and equal to the fit-time representation at full overlap. The score of a row therefore depends only on that row and the frozen model, and is exactly invariant to per-sample scaling. If fewer than model$hp$min_features features are present, the documented neutral score 0 is returned for every row.

Usage

score_dre_ulsif(model, X, meta = NULL)

Arguments

model: A dre_ulsif_model object returned by fit_dre_ulsif.
X: Numeric matrix (samples $\times$ features) of non-negative abundance values. Columns must be named feature ids.
meta: Optional per-sample metadata. Accepted for interface uniformity and ignored by this method.

Value

Plain finite numeric vector of length nrow(X). Larger values are more case-like (the specimen lies where the case density dominates the control density). Scoring uses only each row's own values plus the frozen centres, bandwidth and coefficients.

References

Kanamori T, Hido S, Sugiyama M. (2009) A least-squares approach to direct importance estimation. Journal of Machine Learning Research 10: 1391-1445.

Examples

if (FALSE) { # \dontrun{
set.seed(1)
n <- 120; p <- 30; k <- 10
L <- matrix(stats::rnorm(n * p, 4, 0.5), nrow = n,
            dimnames = list(NULL, paste0("miR-", seq_len(p))))
y <- rep(c(0, 1), each = n / 2)
L[y == 1, seq_len(k)] <- L[y == 1, seq_len(k)] + 1.2
X <- exp(L)
model <- fit_dre_ulsif(X, y)
score_dre_ulsif(model, X[1, , drop = FALSE])
} # }