StandardScaler源码

StandardScaler是scikit-learn中的一个类，用于将数据集进行标准化处理（即将数据集按照特征的均值和方差进行缩放）。下面是StandardScaler的源码：

class StandardScaler(BaseEstimator, TransformerMixin):
    """Standardize features by removing the mean and scaling to unit variance

    The standard score of a sample `x` is calculated as:

        z = (x - u) / s

    where `u` is the mean of the training samples and `s` is the standard
    deviation of the training samples.

    Parameters
    ----------
    copy : bool, default=True
        If False, try to avoid a copy and do inplace scaling instead.
        This is not guaranteed to always work inplace; e.g. if the data is
        not a NumPy array, it may still remain unchanged.

    with_mean : bool, default=True
        If True, center the data before scaling.
        This does not work (and will raise an exception) when attempted on
        sparse matrices.

    with_std : bool, default=True
        If True, scale the data to unit variance (or equivalently,
        unit standard deviation).

    Attributes
    ----------
    scale_ : ndarray of shape (n_features,)
        Per feature relative scaling of the data.

        .. versionadded:: 0.17
           *scale_*

    mean_ : ndarray of shape (n_features,)
        The mean value for each feature in the training set.

    var_ : ndarray of shape (n_features,)
        The variance for each feature in the training set. Used to compute
        `scale_`

    n_samples_seen_ : int or ndarray of shape (n_features,)
        The number of samples processed by the estimator for each feature.
        If there are not missing samples, the n_samples_seen will be an
        integer, otherwise it will be an array with dtype int. By default,
        this will be initialized to zero, and incremented each time a new
        sample is processed. If `sample_weight` is not None, then the
        `n_samples_seen` will be weighted similarly.

        .. versionadded:: 0.24

        .. versionchanged:: 1.0
            `n_samples_seen_` is initialized to zero by default.

    Examples
    --------
    >>> from sklearn.preprocessing import StandardScaler
    >>> import numpy as np
    >>> X = np.array([[-1., -1., 2.],
    ...               [-2., -1., 4.],
    ...               [-3., -2., 1.]])
    >>> scaler = StandardScaler()
    >>> scaler.fit(X)
    StandardScaler()
    >>> print(scaler.mean_)
    [-2.  -1.33333333  2.33333333]
    >>> print(scaler.scale_)
    [0.81649658 0.47140452 1.24721913]
    >>> print(scaler.transform(X))
    [[ 0.26726124  0.          0.12309149]
     [ 0.         0.          1.33630621]
     [-0.26726124 -1.          -1.4593977 ]]
    """

    def __init__(self, *, copy=True, with_mean=True, with_std=True):
        self.with_mean = with_mean
        self.with_std = with_std
        self.copy = copy

    def fit(self, X, y=None):
        """Compute the mean and std to be used for later scaling.

        Parameters
        ----------
        X : {array-like, sparse matrix} of shape (n_samples, n_features)
            The data used to compute the mean and standard deviation
            used for later scaling along the features axis.
        y : None
            Ignored

        Returns
        -------
        self : object
            Returns self.
        """
        if self.with_mean:
            self.mean_ = np.mean(X, axis=0)
        else:
            self.mean_ = None
        if self.with_std:
            self.var_ = np.var(X, axis=0)
            self.scale_ = np.sqrt(self.var_)
        else:
            self.var_ = None
            self.scale_ = None
        self.n_samples_seen_ = _num_samples(X)
        return self

    def transform(self, X, y='deprecated', copy=None):
        """Perform standardization by centering and scaling

        Parameters
        ----------
        X : {array-like, sparse matrix} of shape (n_samples, n_features)
            The data to transform, row by row.
            If a sparse matrix is provided, it will be converted into
            a sparse ``csr_matrix``. **However, CSC or COO matrices are
            not supported yet**.

        y : (ignored)
            .. deprecated:: 0.24
               `y` parameter is deprecated and will be removed in v1.1
               Use sklearn.compose.TransformedTargetRegressor instead.

        copy : bool, default=None
            Copy the input X or not.

        Returns
        -------
        X_tr : {ndarray, sparse matrix} of shape (n_samples, n_features)
            Transformed array.

        """
        check_is_fitted(self)

        if isinstance(X, (SparseArray, spmatrix)):
            if self.with_mean:
                raise ValueError(
                    "Cannot center sparse matrices: use `with_mean=False` "
                    "instead. See docstring for motivation and alternatives.")
            if self.scale_ is not None:
                inplace_column_scale(X, 1 / self.scale_)
        else:
            X = check_array(X, accept_sparse='csc', copy=copy,
                            estimator=self, dtype=FLOAT_DTYPES,
                            force_all_finite='allow-nan')

            if self.with_mean:
                X -= self.mean_
            if self.with_std:
                X /= self.scale_
        return X

    def inverse_transform(self, X, copy=None):
        """Scale back the data to the original representation

        Parameters
        ----------
        X : {ndarray, sparse matrix} of shape (n_samples, n_features)
            The data to transform back, row by row.

        copy : bool, default=None
            Copy the input X or not.

        Returns
        -------
        X_original : {ndarray, sparse matrix} of shape (n_samples, n_features)
            Transformed data.
        """
        check_is_fitted(self)
        if isinstance(X, (SparseArray, spmatrix)):
            if self.with_mean:
                raise ValueError(
                    "Cannot center sparse matrices: use `with_mean=False` "
                    "instead. See docstring for motivation and alternatives.")
            if self.scale_ is not None:
                inplace_column_scale(X, self.scale_)
        else:
            X = check_array(X, accept_sparse='csc', copy=copy,
                            estimator=self, dtype=FLOAT_DTYPES,
                            force_all_finite='allow-nan')

            if self.with_std:
                X *= self.scale_
            if self.with_mean:
                X += self.mean_
        return X

    def _more_tags(self):
        return {'allow_nan': True}

在上面的代码中，fit()方法用于计算均值和标准差，transform()方法用于将数据集进行标准化处理，inverse_transform()方法用于将标准化后的数据集进行还原，_more_tags()方法是scikit-learn的内部方法，用于提供更多的元数据信息。注意，StandardScaler会对每个特征进行标准化处理，因此它只能应用于连续型特征，而不能应用于类别型特征。