lin_fit

Signature

lin_fit(
    x: ArrayLike,
    y: ArrayLike,
    sigma_y: ArrayLike,
    *,
    sigma_x: ArrayLike | None = None,
    tol: float = 1e-10,
    max_iter: int = 100,
    style: str | None = "mespy",
    xlabel: str = "x [xu]",
    ylabel: str = "y [uy]",
    residuals_label: str = "Residuals",
    normalize_residuals: bool = False,
    band_label: str = r"$\pm 1 \sigma$ retta",
    fit_label: str = r"Fit",
    title: str | None = None,
    decimals: int = 3,
    show_plot: bool = True,
    show_band: bool = True,
    show_legend: bool = True,
    show_fit_params: bool = False,
    show_grid: bool = True,
    xlim: ArrayLike | None = None,
    ylim: ArrayLike | None = None,
    figsize: ArrayLike | None = None,
    dpi: int | None = None,
    save_path: str | None = None,
    title_fontsize: int | float | None = None,
    title_pad: int | float | None = None,
    legend_fontsize: int | float | None = None,
    legend_loc: str | None = None,
    point_color: str | None = None,
    fit_color: str | None = None,
    band_color: str | None = None,
    res_line_color: str | None = None,
    data_alpha: float = 1.0,
    band_alpha: float = 0.20,
    grid_alpha: float | None = None,
) -> LinearFitResult

lin_fit performs a weighted linear fit y = m x + c on experimental data, with optional support for uncertainties on x as well.

By default it applies the mespy Matplotlib style; pass style=None to use the current rcParams, or the name of another style to delegate to Matplotlib.

I parametri figsize, dpi, title_fontsize, title_pad, legend_fontsize, legend_loc, point_color, fit_color, band_color, res_line_color e grid_alpha possono ereditare lo stile attivo quando lasciati a None. I parametri data_alpha e band_alpha restano invece override espliciti della singola chiamata.

fit_label e band_label permettono di personalizzare i testi della legenda di retta e banda. In particolare, fit_label viene usato solo quando show_fit_params=False; con show_fit_params=True la funzione genera automaticamente una label con m e c.

normalize_residuals controlla solo il pannello inferiore del grafico. Con il default False, il pannello mostra i residui fisici y_i - (m x_i + c). Con True, mostra invece i residui normalizzati r_i / sigma_eff_i, usando la stessa varianza efficace impiegata per chi2. Il campo LinearFitResult.residuals resta sempre non normalizzato.

This page remains the quick access point for the function. Practical details are collected in Overview, while How it works is prepared as a separate subpage.

On this page