Score ordinal Local Binary Pattern (inv-olbp) within-sample discriminator

Scores each row of X independently with the frozen model learned by fit_inv_olbp. The model-universe features present in X are taken in the frozen canonical order, and for one specimen its within-sample ranks over exactly those present features are turned into the ordinal Local Binary Pattern histogram (same neighbour offsets, same bit packing, same $2^P$ bins as at fitting; the circular neighbour wrap is recomputed on the present length $M$). The histogram is standardised by the frozen center/scale and passed through the frozen linear head. Larger values are more case-like.

Scoring uses only each row's own values plus the frozen model – no test-batch renormalisation, no cross-row coupling, no statistic estimated from X. If the number of present model-universe features is below model$hp$min_features or not greater than the neighbour count $P$ (so a circular OLBP code is undefined), the documented neutral score 0 is returned for every row.

Usage

score_inv_olbp(model, X, meta = NULL)

Arguments

model: An inv_olbp_model object returned by fit_inv_olbp.
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 score of a row depends only on that row's values and the frozen model, and is exactly invariant to per-sample positive scaling.

References

Ojala T, Pietikainen M, Maenpaa T. (2002) Multiresolution gray-scale and rotation invariant texture classification with local binary patterns. IEEE Transactions on Pattern Analysis and Machine Intelligence 24(7): 971-987.

Examples

if (FALSE) { # \dontrun{
set.seed(1)
p <- 40
X <- matrix(stats::rgamma(120 * p, shape = 20, rate = 0.2), nrow = 120,
            dimnames = list(NULL, paste0("miR-", sprintf("%03d", seq_len(p)))))
y <- rep(c(0, 1), each = 60)
X[y == 1, 11:16] <- X[y == 1, 11:16] + 150
model <- fit_inv_olbp(X, y)
score_inv_olbp(model, X[1, , drop = FALSE])
} # }