Score Mondrian-conformal class-conditional LRT discriminator — score_conf

Scores each row of X independently with a frozen fit_conf_mondrian model. A specimen is transformed to within-sample rCLR coordinates over the model features present in X; its class-specific nonconformity scores are compared with the frozen class-conditional calibration scores; and the returned score is $$\log p_{case}(x) - \log p_{control}(x).$$

Usage

score_conf_mondrian(model, X, meta = NULL)

Arguments

model: A conf_mondrian_model object returned by fit_conf_mondrian.
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. If fewer than model$hp$min_features model features are present, the documented neutral score 0 is returned for every row.

Details

At scoring time the present feature set is intersect(model$feature_universe, colnames(X)). Class centroids, diagonal variances, and calibration nonconformity scores are re-derived over that present subset from the frozen raw anchors only. Each row's score then uses only that row's own rCLR vector plus the frozen model-derived representation, so no scored-batch statistic or cross-row coupling is used.

References

Vovk V, Gammerman A, Shafer G. (2005) Algorithmic Learning in a Random World. Springer.

Shafer G, Vovk V. (2008) A tutorial on conformal prediction. Journal of Machine Learning Research 9: 371-421.

Examples

if (FALSE) { # \dontrun{
set.seed(1)
X <- matrix(stats::rgamma(60 * 30, shape = 2), nrow = 60,
            dimnames = list(NULL, paste0("miR-", seq_len(30))))
y <- rep(c(0, 1), each = 30)
X[y == 1, 1:5] <- X[y == 1, 1:5] + 20
X[y == 0, 6:10] <- X[y == 0, 6:10] + 20
model <- fit_conf_mondrian(X, y)
score_conf_mondrian(model, X[1, , drop = FALSE])
} # }