python基础技巧(四)——sklearn

• 激活函数

tf.sigmoid(x)
tf.tanh(x)
tf.softsign(x)
tf.nn.relu(x)
tf.nn.elu(x)
tf.nn.bias_add(values,bias)

• 损失优化方法
  • tf.train.

tf.train.GradientDescentOptimizer(learning_rate=,use_locking=,name=)
tf.train.AdadeltaOptimizer(learning_rate=,use_locking=,name=)
tf.train.AdagradOptimizer(learning_rate=,use_locking=,name=)
tf.train.AdamOptimizer(learning_rate=,use_locking=,name=)
...

• preprocessing

  https://blog.csdn.net/csmqq/article/details/51461696

数据预处理——标准化，均值去除和按方差比例缩放（Standardization, or mean removal and variance scaling）

• StandardScaler计算训练集的平均值和标准差，以便测试数据集使用相同的变换。

• scale 零均值单位方差

  • 1）若设置with_mean=False 或者 with_std=False，则不做centering 或者scaling处理。

    2）scale和StandardScaler可以用于回归模型中的目标值处理。

• MinMaxScaler(最小最大值标准化)

• MaxAbsScaler（绝对值最大标准化）

• metrics

  https://www.cnblogs.com/harvey888/p/6964741.html

各种误差

from sklearn import metrics
metrics.mean_absolute_error(Y_predict,Y_true)
metrics.mean_squared_error(Y_predict,Y_true)
metrics.mean_squared_log_error(Y_predict,Y_true)
metrics.median_absolute_error(Y_predict,Y_true)
...

• cross_validation

  将数据集拆分为训练集和测试集

  • train_size=0.7 表示训练集占全集的70%
  • test_size=0.2 表示测数据占全集的20%

cross_validation.train_test_split(X,y,train_size=0.7,test_size=0.2)

• sklearn.utils.shuffle

随机排序

*arrays : sequence of indexable data-structures

Indexable data-structures can be arrays, lists, dataframes or scipy sparse matrices with consistent first dimension.

random_state : int, RandomState instance or None, optional (default=None)

The seed of the pseudo random number generator to use when shuffling the data. If int, random_state is the seed used by the random number generator; If RandomState instance, random_state is the random number generator; If None, the random number generator is the RandomState instance used by np.random.

n_samples : int, None by default

Number of samples to generate. If left to None this is automatically set to the first dimension of the arrays.