yobx.sklearn.decomposition.truncated_svd

yobx.sklearn.decomposition.truncated_svd.sklearn_truncated_svd(g: GraphBuilderExtendedProtocol, sts: Dict, outputs: List[str], estimator: TruncatedSVD, X: str, name: str = 'truncated_svd') → str

Converts a sklearn.decomposition.TruncatedSVD into ONNX.

Unlike PCA, TruncatedSVD does not center the data before projecting. The transformation is simply a matrix multiplication:

X  ──MatMul(components_.T)──►  output

Parameters:

• g – the graph builder to add nodes to

• sts – shapes defined by scikit-learn

• estimator – a fitted TruncatedSVD

• outputs – desired output names (projected inputs)

• X – input tensor name

• name – prefix name for the added nodes

Returns:

output tensor name