学习多标签分类

Question

我尝试使用从nolearn包导入的DBN函数，下面是我的代码:

from nolearn.dbn import DBN
import numpy as np
from sklearn import cross_validation
fileName = 'data.csv'
fileName_1 = 'label.csv'
data = np.genfromtxt(fileName, dtype=float, delimiter = ',')
label = np.genfromtxt(fileName_1, dtype=int, delimiter = ',')
clf = DBN(
    [data, 300, 10],
    learn_rates=0.3,
    learn_rate_decays=0.9,
    epochs=10,
    verbose=1,
    )
clf.fit(data,label)
score = cross_validation.cross_val_score(clf, data, label,scoring='f1', cv=10)
print score

由于我的数据具有形状(1231,229)和形状(1231,13)的标签，因此标签集看起来像([0 0 1 0 1 0 0 0 0 1 1 0 0]…，[....])，当我运行代码时，我得到了这个错误消息:错误的输入形状(1231,13)。我想这里可能发生了两个问题:

1. DBN不支持多标签分类
2. my label不适合用于DBN拟合函数

Answer 1

正如Francisco Vargas所提到的，nolearn.dbn已被弃用，您应该使用nolearn.lasagne(如果可以的话)。

如果您想在千层面中进行多标签分类，那么您应该将regression参数设置为True，定义验证分数和自定义损失。

下面是一个例子:

import numpy as np
import theano.tensor as T
from lasagne import layers
from lasagne.updates import nesterov_momentum
from nolearn.lasagne import NeuralNet
from nolearn.lasagne import BatchIterator
from lasagne import nonlinearities
# custom loss: multi label cross entropy
def multilabel_objective(predictions, targets):
    epsilon = np.float32(1.0e-6)
    one = np.float32(1.0)
    pred = T.clip(predictions, epsilon, one - epsilon)
    return -T.sum(targets * T.log(pred) + (one - targets) * T.log(one - pred), axis=1)

net = NeuralNet(
    # customize "layers" to represent the architecture you want
    # here I took a dummy architecture
    layers=[(layers.InputLayer, {"name": 'input', 'shape': (None, 1, 229, 1)}),
            (layers.DenseLayer, {"name": 'hidden1', 'num_units': 20}),
            (layers.DenseLayer, {"name": 'output', 'nonlinearity': nonlinearities.sigmoid, 'num_units': 13})], #because you have 13 outputs
    # optimization method:
    update=nesterov_momentum,
    update_learning_rate=5*10**(-3),
    update_momentum=0.9,
    max_epochs=500,  # we want to train this many epochs
    verbose=1,
    #Here are the important parameters for multi labels
    regression=True,  
    objective_loss_function=multilabel_objective,
    custom_score=("validation score", lambda x, y: np.mean(np.abs(x - y)))
    )
net.fit(X_train, labels_train)

Answer 2

Fit调用BuildDBN，可以在这里找到，需要注意的重要事情是dbn已被弃用，您只能找到它old_commits。无论如何，如果你正在寻找额外的信息，从我在你的代码片段中看到的检查这两个可能是好的，DBN即[data, 300, 10]的第一个参数应该是基于文档和源代码的[data.shape[1], 300, 10]。