机器学习正则化公式及python代码实现

导入模型：

from sklearn.preprocessing import Normalizer

sklearn.preprocessing.Normalizer(norm=’l2’, copy=True)

其中norm=l1,l2,或max

当norm=l1时，样本各个特征值除以各个特征值的绝对值之和

当norm=l2时，样本各个特征值除以各个特征值的平方之和的2次幂

当norm=max时，样本各个特征值除以样本中特征值最大的值

from sklearn.preprocessing import Normalizer
A = [[1,2,2],[3,4,-1],[2,1,-1]]
scaler = Normalizer(norm = 'l2')
scaler  =scaler.fit(A)
result = scaler.transform(A)
print(result)

当norm为l2时

 当norm=l1时

from sklearn.preprocessing import Normalizer
A = [[1,2,2],[3,4,-1],[2,1,-1]]
scaler = Normalizer(norm = 'l1')
scaler1  =scaler.fit(A)
result = scaler1.transform(A)
print(result)

 

 当norm = max时

from sklearn.preprocessing import Normalizer
A = [[1,2,2],[3,4,-1],[2,1,-1]]
scaler = Normalizer(norm = 'max')
scaler1  =scaler.fit(A)
result = scaler1.transform(A)
print(result)