scikit-learn linearRegression 1.1.2 岭回归

Ridge 岭回归通过对回归稀疏增加罚项来解决 普通最小二乘法 的一些问题.岭回归系数通过最小化带罚项的残差平方和

上述公式中,  是控制模型复杂度的因子(可看做收缩率的大小) :  越大，收缩率越大，那么系数对于共线性的鲁棒性更强

一、一般线性回归遇到的问题

    在处理复杂的数据的回归问题时，普通的线性回归会遇到一些问题，主要表现在：

• 预测精度：这里要处理好这样一对为题，即样本的数量和特征的数量
  • 时，最小二乘回归会有较小的方差
    
  • 时，容易产生过拟合
    
  • 时，最小二乘回归得不到有意义的结果
    
• 模型的解释能力：如果模型中的特征之间有相互关系，这样会增加模型的复杂程度，并且对整个模型的解释能力并没有提高，这时，我们就要进行特征选择。

以上的这些问题，主要就是表现在模型的方差和偏差问题上，这样的关系可以通过下图说明：

（摘自：机器学习实战）

方差指的是模型之间的差异，而偏差指的是模型预测值和数据之间的差异。我们需要找到方差和偏差的折中。

二、岭回归的概念

    在进行特征选择时，一般有三种方式：

• 子集选择
• 收缩方式(Shrinkage method)，又称为正则化(Regularization)。主要包括岭回归和lasso回归。
• 维数缩减

    岭回归(Ridge Regression)是在平方误差的基础上增加正则项

,

通过确定的值可以使得在方差和偏差之间达到平衡：随着的增大，模型方差减小而偏差增大。

    对求导，结果为

令其为0，可求得的值：

三、实验的过程

    我们去探讨一下取不同的对整个模型的影响。

和其他线性模型一样，Ridge 调用 fit 方法，参数为X,y,并且将线性模型拟合的系数  存到成员变量 coef_中。:

>>> from sklearn import linear_model>>> clf = linear_model.Ridge (alpha = .5)>>> clf.fit ([[0, 0], [0, 0], [1, 1]], [0, .1, 1]) Ridge(alpha=0.5, copy_X=True, fit_intercept=True, max_iter=None,      normalize=False, random_state=None, solver='auto', tol=0.001)>>> clf.coef_array([ 0.34545455,  0.34545455])>>> clf.intercept_ 0.13636...

绘制岭系数正则化的函数

岭回归是本例中使用的估计量，美中衍射代表不同系数向量特征，在路径末端，alpha趋于零，解趋于普通最小二乘法，系数表现出现很大的震荡

# Author: Fabian Pedregosa -- <[e-mail]># License: BSD 3 clauseprint(__doc__)import numpy as npimport matplotlib.pyplot as pltfrom sklearn import linear_model# X is the 10x10 Hilbert matrix 希尔伯特矩阵X = 1. / (np.arange(1, 11) + np.arange(0, 10)[:, np.newaxis])y = np.ones(10)################################################################################ Compute paths a为alpha 正则化系数，通过便利-10 到 -2之间200个系数，来寻找最佳系数n_alphas = 200alphas = np.logspace(-10, -2, n_alphas)clf = linear_model.Ridge(fit_intercept=False)coefs = []for a in alphas:    clf.set_params(alpha=a)    clf.fit(X, y)    coefs.append(clf.coef_)################################################################################ Display resultsax = plt.gca()ax.set_color_cycle(['b', 'r', 'g', 'c', 'k', 'y', 'm'])ax.plot(alphas, coefs)ax.set_xscale('log')ax.set_xlim(ax.get_xlim()[::-1])  # reverse axisplt.xlabel('alpha')plt.ylabel('weights')plt.title('Ridge coefficients as a function of the regularization')plt.axis('tight')plt.show()

class sklearn.linear_model.RidgeClassifier(alpha=1.0, fit_intercept=True, normalize=False, copy_X=True, max_iter=None, tol=0.001, class_weight=None, solver='auto', random_state=None)[source]
Classifier using Ridge regression.
Read more in the User Guide.
Parameters:
alpha : float
Regularization strength; must be a positive float. Regularization improves the conditioning of the problem and reduces the variance of the estimates. Larger values specify stronger regularization. Alpha corresponds to C^-1 in other linear models such as LogisticRegression or LinearSVC.
class_weight : dict or ‘balanced’, optional
Weights associated with classes in the form {class_label: weight}. If not given, all classes are supposed to have weight one.
The “balanced” mode uses the values of y to automatically adjust weights inversely proportional to class frequencies in the input data as n_samples / (n_classes * np.bincount(y))
copy_X : boolean, optional, default True
If True, X will be copied; else, it may be overwritten.
fit_intercept : boolean
Whether to calculate the intercept for this model. If set to false, no intercept will be used in calculations (e.g. data is expected to be already centered).
max_iter : int, optional
Maximum number of iterations for conjugate gradient solver. The default value is determined by scipy.sparse.linalg.
normalize : boolean, optional, default False
If True, the regressors X will be normalized before regression. This parameter is ignored when fit_intercept is set to False. When the regressors are normalized, note that this makes the hyperparameters learnt more robust and almost independent of the number of samples. The same property is not valid for standardized data. However, if you wish to standardize, please use preprocessing.StandardScaler before calling fit on an estimator with normalize=False.
solver : {‘auto’, ‘svd’, ‘cholesky’, ‘lsqr’, ‘sparse_cg’, ‘sag’}
Solver to use in the computational routines:
‘auto’ chooses the solver automatically based on the type of data.
‘svd’ uses a Singular Value Decomposition of X to compute the Ridge coefficients. More stable for singular matrices than ‘cholesky’.
‘cholesky’ uses the standard scipy.linalg.solve function to obtain a closed-form solution.
‘sparse_cg’ uses the conjugate gradient solver as found in scipy.sparse.linalg.cg. As an iterative algorithm, this solver is more appropriate than ‘cholesky’ for large-scale data (possibility to set tol and max_iter).
‘lsqr’ uses the dedicated regularized least-squares routine scipy.sparse.linalg.lsqr. It is the fastest but may not be available in old scipy versions. It also uses an iterative procedure.
‘sag’ uses a Stochastic Average Gradient descent. It also uses an iterative procedure, and is faster than other solvers when both n_samples and n_features are large.
New in version 0.17: Stochastic Average Gradient descent solver.
tol : float
Precision of the solution.
random_state : int seed, RandomState instance, or None (default)
The seed of the pseudo random number generator to use when shuffling the data. Used in ‘sag’ solver.
Attributes:
coef_ : array, shape (n_features,) or (n_classes, n_features)
Weight vector(s).
intercept_ : float | array, shape = (n_targets,)
Independent term in decision function. Set to 0.0 if fit_intercept = False.
n_iter_ : array or None, shape (n_targets,)
Actual number of iterations for each target. Available only for sag and lsqr solvers. Other solvers will return None.
See also
 
Ridge, RidgeClassifierCV
Notes
For multi-class classification, n_class classifiers are trained in a one-versus-all approach. Concretely, this is implemented by taking advantage of the multi-variate response support in Ridge.
Methods
decision_function(X)Predict confidence scores for samples.fit(X, y[, sample_weight])Fit Ridge regression model.get_params([deep])Get parameters for this estimator.predict(X)Predict class labels for samples in X.score(X, y[, sample_weight])Returns the mean accuracy on the given test data and labels.set_params(\*\*params)Set the parameters of this estimator.
__init__(alpha=1.0, fit_intercept=True, normalize=False, copy_X=True, max_iter=None, tol=0.001, class_weight=None, solver='auto', random_state=None)[source]
decision_function(X)[source]
Predict confidence scores for samples.
The confidence score for a sample is the signed distance of that sample to the hyperplane.
Parameters:
X : {array-like, sparse matrix}, shape = (n_samples, n_features)
Samples.
Returns:
array, shape=(n_samples,) if n_classes == 2 else (n_samples, n_classes) :
Confidence scores per (sample, class) combination. In the binary case, confidence score for self.classes_[1] where >0 means this class would be predicted.
fit(X, y, sample_weight=None)[source]
Fit Ridge regression model.
Parameters:
X : {array-like, sparse matrix}, shape = [n_samples,n_features]
Training data
y : array-like, shape = [n_samples]
Target values
sample_weight : float or numpy array of shape (n_samples,)
Sample weight.
New in version 0.17: sample_weight support to Classifier.
Returns:
self : returns an instance of self.
get_params(deep=True)[source]
Get parameters for this estimator.
Parameters:
deep : boolean, optional
If True, will return the parameters for this estimator and contained subobjects that are estimators.
Returns:
params : mapping of string to any
Parameter names mapped to their values.
predict(X)[source]
Predict class labels for samples in X.
Parameters:
X : {array-like, sparse matrix}, shape = [n_samples, n_features]
Samples.
Returns:
C : array, shape = [n_samples]
Predicted class label per sample.
score(X, y, sample_weight=None)[source]
Returns the mean accuracy on the given test data and labels.
In multi-label classification, this is the subset accuracy which is a harsh metric since you require for each sample that each label set be correctly predicted.
Parameters:
X : array-like, shape = (n_samples, n_features)
Test samples.
y : array-like, shape = (n_samples) or (n_samples, n_outputs)
True labels for X.
sample_weight : array-like, shape = [n_samples], optional
Sample weights.
Returns:
score : float
Mean accuracy of self.predict(X) wrt. y.
set_params(**params)[source]
Set the parameters of this estimator.
The method works on simple estimators as well as on nested objects (such as pipelines). The latter have parameters of the form <component>__<parameter> so that it’s possible to update each component of a nested object.

文章参考：http://blog.csdn.net/google19890102/article/details/27228279