model.py 文件源码

def ols_coefficients_to_dataframe(coefs):
    """
    Take a series containing OLS coefficients and convert it
    to a data frame.

    Parameters
    ----------
    coefs : pandas Series
        Series with feature names in the index and the coefficient
        values as the data, obtained from a linear model trained
        using `statsmodels.OLS`.

    Returns
    -------
    df_coef : pandas DataFrame
        Data frame with two columns, the first being the feature name
        and the second being the coefficient value.

    Note
    ----
    The first row in the output data frame is always for the intercept
    and the rest are sorted by feature name.
    """
    # first create a sorted data frame for all the non-intercept features
    non_intercept_columns = [c for c in coefs.index if c != 'const']
    df_non_intercept = pd.DataFrame(coefs.filter(non_intercept_columns), columns=['coefficient'])
    df_non_intercept.index.name = 'feature'
    df_non_intercept = df_non_intercept.sort_index()
    df_non_intercept.reset_index(inplace=True)

    # now create a data frame that just has the intercept
    df_intercept = pd.DataFrame([{'feature': 'Intercept',
                                  'coefficient': coefs['const']}])

    # append the non-intercept frame to the intercept one
    df_coef = df_intercept.append(df_non_intercept, ignore_index=True)

    # we always want to have the feature column first
    df_coef = df_coef[['feature', 'coefficient']]

    return df_coef