LogisticRegression — transformers/LogisticFunction.py

Logistic accuracy contribution transformer. Computes the per-record accuracy contribution of a logistic classifier — i.e., whether the model's binary prediction matches the true label — normalised by the total sample size N.

eta = x @ coefs^T pred = (eta >= 0) ? 1 : 0 result = (pred == y) ? 1/N : 0 SLT_i = Concat(result, TS_i)

When summed across all records (via the Adder aggregator), the SMT root value equals the overall classification accuracy.

Constructor

python LogisticRegression(size=0, length_transform_salt=0, beta=None, sample_size=None)

Parameter	Type	Default	Description
size	int	0	Expected output shape.
length_transform_salt	int	0	Dimensionality of TS_i.
beta	torch.Tensor	None	Coefficient vector of the logistic model, shape (n_coefs, 1).
sample_size	torch.Tensor	None	Total sample size N, shape (1, 1). Sourced from the root value of the Length SMT.

forward(a, salt, beta, sample_size)

Parameter	Type	Description
a	torch.Tensor	Salted raw value SD_i. First x_dim columns are features x; column x_dim is label y.
salt	torch.Tensor	Transform salt TS_i.
beta	torch.Tensor	Logistic coefficient vector.
sample_size	torch.Tensor	N as a (1, 1) tensor.

Returns torch.Tensor — Concat((pred == y) / N, salt).

Note

beta and sample_size are private circuit inputs included in the ONNX export — the prover keeps the model coefficients secret.

---

See index.md for the shared interface inherited by all transformers.