quantlab.model.predefined.linear.gram

quantlab.model.predefined.linear.gram(x, y, block_rows=262144)

Return (X'X, X'y, sum(X), sum(y), n) in float64, accumulated by blocks of rows.

Examples

>>> xtx, xty, sx, sy, n = gram(np.eye(3, dtype=np.float32), np.arange(3.0), block_rows=2)
>>> xtx.diagonal().tolist(), xty.tolist(), n
([1.0, 1.0, 1.0], [0.0, 1.0, 2.0], 3)