Function custom_rmse_model

demo/guide-python/multioutput_regression.py:69–120

Train using Python implementation of Squared Error.

(strategy: str, ax: Optional[matplotlib.axes.Axes])

def custom_rmse_model(strategy: str, ax: Optional[matplotlib.axes.Axes]) -> None:
    """Train using Python implementation of Squared Error."""

    def gradient(predt: np.ndarray, dtrain: xgb.DMatrix) -> np.ndarray:
        """Compute the gradient squared error."""
        y = dtrain.get_label().reshape(predt.shape)
        return predt - y

    def hessian(predt: np.ndarray, dtrain: xgb.DMatrix) -> np.ndarray:
        """Compute the hessian for squared error."""
        return np.ones(predt.shape)

    def squared_log(
        predt: np.ndarray, dtrain: xgb.DMatrix
    ) -> Tuple[np.ndarray, np.ndarray]:
        grad = gradient(predt, dtrain)
        hess = hessian(predt, dtrain)
        # both numpy.ndarray and cupy.ndarray works.
        return grad, hess

    def rmse(predt: np.ndarray, dtrain: xgb.DMatrix) -> Tuple[str, float]:
        y = dtrain.get_label().reshape(predt.shape)
        v = np.sqrt(np.mean(np.power(y - predt, 2)))
        return "PyRMSE", v

    X, y = gen_circle()
    Xy = xgb.DMatrix(X, y)
    results: Dict[str, Dict[str, List[float]]] = {}
    # Make sure the `num_target` is passed to XGBoost when custom objective is used.
    # When builtin objective is used, XGBoost can figure out the number of targets
    # automatically.
    booster = xgb.train(
        {
            "tree_method": "hist",
            "num_target": y.shape[1],
            "multi_strategy": strategy,
        },
        dtrain=Xy,
        num_boost_round=128,
        obj=squared_log,
        evals=[(Xy, "Train")],
        evals_result=results,
        custom_metric=rmse,
    )

    y_predt = booster.inplace_predict(X)
    if ax:
        plot_predt(y, y_predt, f"PyRMSE-{strategy}", ax)

    np.testing.assert_allclose(
        results["Train"]["rmse"], results["Train"]["PyRMSE"], rtol=1e-2
    )


if __name__ == "__main__":