Score kernel mean-embedding witness-at-a-point discriminator

Scores each row of X independently with the frozen model learned by fit_kme_witness. For one specimen \(x\), the score is the witness function evaluated at that point: $$mean_i k(phi(x), phi(a_i^{case})) - mean_j k(phi(x), phi(a_j^{control})),$$ where \(k(a,b) = exp(-gamma ||a-b||^2)\). At scoring time, the present feature set is intersect(model$feature_universe, colnames(X)). Both the specimen and the frozen raw anchors are closed to proportions over exactly that present set and then centred-log-ratio transformed, which keeps partial-overlap distances consistent (and the score exactly invariant to per-sample scaling) without using any scored-batch statistic. The frozen bandwidth model$gamma is applied unchanged in this reduced present-feature space; it is not re-tuned to the overlap, which is what keeps scoring single-sample. Re-deriving the anchor representation over present uses only frozen training data, so the score depends only on the specimen and the frozen model. If fewer than model$hp$min_features features are present, the documented neutral score 0 is returned for every row.

Usage

score_kme_witness(model, X, meta = NULL)

Arguments

model

A kme_witness_model object returned by fit_kme_witness.

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 because the witness is case mean embedding minus control mean embedding. Scoring uses only each row's own values plus the frozen anchors, gamma, and pc.

References

Gretton A, Borgwardt KM, Rasch MJ, Scholkopf B, Smola A. (2012) A kernel two-sample test. Journal of Machine Learning Research 13: 723-773.

Sutherland DJ, Tung HY, Strathmann H, De S, Ramdas A, Smola A, Gretton A. (2017) Generative models and model criticism via optimized maximum mean discrepancy. International Conference on Learning Representations.

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_kme_witness(X, y)
score_kme_witness(model, X[1, , drop = FALSE])
} # }