sklearn: LogisticRegression - predict_proba(X) - 计算

sklearn: LogisticRegression - predict_proba(X) - calculation

我想知道是否有人可以快速查看以下代码片段并指出我在计算模型中每个 class 和我的样本概率时的误解相关代码错误。我试图手动计算 sklearn 函数 lm.predict_proba(X) 提供的结果,遗憾的是结果不同,所以我犯了一个错误。

我认为该错误将出现在以下代码演练的 "d" 部分。也许在数学上,但我不明白为什么。

a) 创建和训练逻辑回归模型(工作正常)

lm = LogisticRegression(random_state=413, multi_class='multinomial', solver='newton-cg')
lm.fit(X, train_labels)

b) 保存系数和偏差(工作正常)

W = lm.coef_
b = lm.intercept_

c) 使用 lm.predict_proba(X)(工作正常)

def reshape_single_element(x,num):
    singleElement = x[num]
    nx,ny = singleElement.shape
    return  singleElement.reshape((1,nx*ny))

select_image_number = 6 
X_select_image_data=reshape_single_element(train_dataset,select_image_number)
Y_probabilities =  lm.predict_proba(X_select_image_data)
Y_pandas_probabilities = pd.Series(Y_probabilities[0], index=['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j'])
print"estimate probabilities for each class: \n" ,Y_pandas_probabilities , "\n"
print"all probabilities by lm.predict_proba(..) sum up to ", np.sum(Y_probabilities) , "\n"

输出为:

estimate probabilities for each class: 
a 0.595426
b 0.019244
c 0.001343
d 0.004033
e 0.017185
f 0.004193
g 0.160380
h 0.158245
i 0.003093
j 0.036860
dtype: float64 
all probabilities by lm.predict_proba(..) sum up to 1.0

d) 手动执行由 lm.predict_proba 完成的计算(没有 error/warning,但结果不一样)

manual_calculated_probabilities = []
for select_class_k in range(0,10):  #a=0. b=1, c=3 ...
    z_for_class_k = (np.sum(W[select_class_k] *X_select_image_data) + b[select_class_k] )
    p_for_class_k = 1/ (1 + math.exp(-z_for_class_k))
    manual_calculated_probabilities.append(p_for_class_k)

print "formula: ", manual_calculated_probabilities , "\n"

def softmax(x):
    """Compute softmax values for each sets of scores in x."""
    e = np.exp(x)
    dist = e / np.sum(np.exp(x),axis=0)
    return dist

abc = softmax(manual_calculated_probabilities)
print "softmax:" , abc

输出为:

formula: [0.9667598370531315, 0.48453459121301334, 0.06154496922245115, 0.16456194859398865, 0.45634781280053394, 0.16999340794727547, 0.8867996361191054, 0.8854473986336552, 0.13124464656251109, 0.642913996162282]

softmax: [ 0.15329642 0.09464644 0.0620015 0.0687293 0.0920159 0.069103610.14151607 0.14132483 0.06647715 0.11088877]

由于 github logistic.py

的评论,使用了 Softmax
For a multi_class problem, if multi_class is set to be "multinomial" the softmax function is used to find the predicted probability of each class.

注:

print "shape of X: " , X_select_image_data.shape
print "shape of W: " , W.shape
print "shape of b: " , b.shape

shape of X:  (1, 784)
shape of W:  (10, 784)
shape of b:  (10,)

我发现了一个非常相似的问题 here,但遗憾的是我无法将其调整到我的代码中,所以预测结果是一样的。我尝试了许多不同的组合来计算变量 'z_for_class_k' 和 'p_for_class_k' 但遗憾的是没有成功地从 'predict_proba(X)'.

重现预测值

我认为问题出在

p_for_class_k = 1/ (1 + math.exp(-z_for_class_k))

1 / (1 + exp(-logit)) 是一种仅适用于二进制问题的简化。

真正的等式在简化之前是这样的:

p_for_classA = exp(logit_classA) / [1 + exp(logit_classA) + exp(logit_classB) ... + exp(logit_classC)]

换句话说,在计算特定 class 的概率时,您必须将其他 class 的所有权重和偏差也纳入您的公式。

我没有数据来对此进行测试,但希望这能为您指明正确的方向。

改变

p_for_class_k = 1/ (1 + math.exp(-z_for_class_k))
manual_calculated_probabilities.append(p_for_class_k)

manual_calculated_probabilities.append(z_for_class_k)

也就是 softmax 的输入在您的符号中是“z”而不是“p”。 multinomial logistic

我能够通过执行以下操作来复制方法 lr.predict_proba

>>> sigmoid = lambda x: 1/(1+np.exp(-x))
>>> sigmoid(lr.intercept_+np.sum(lr.coef_*X.values, axis=1))

假设 X 是一个 numpy 数组,lr 是一个来自 sklearn 的对象。