python - class_weights 如何应用于 sklearn 逻辑回归？

Question

我对 sklearn 如何应用我们提供的类权重感兴趣。 documentation没有明确说明类权重的应用位置和方式。阅读源代码也没有帮助(似乎 sklearn.svm.liblinear 用于优化，我无法阅读源代码，因为它是一个 .pyd 文件...)

但我猜它适用于成本函数:指定类别权重时，相应类别的成本将乘以类别权重。例如，如果我分别从 0 类(权重=0.5)和 1 类(权重=1)分别有 2 个观察值，那么成本函数将是:

Cost = 0.5*log(...X_0,y_0...) + 1*log(...X_1,y_1...) + penalization

有谁知道这是否正确？

Answer 1

检查 the following lines in the source code :

le = LabelEncoder()
if isinstance(class_weight, dict) or multi_class == 'multinomial':
    class_weight_ = compute_class_weight(class_weight, classes, y)
    sample_weight *= class_weight_[le.fit_transform(y)]

Here is the source code for the compute_class_weight() function :

...
else:
    # user-defined dictionary
    weight = np.ones(classes.shape[0], dtype=np.float64, order='C')
    if not isinstance(class_weight, dict):
        raise ValueError("class_weight must be dict, 'balanced', or None,"
                         " got: %r" % class_weight)
    for c in class_weight:
        i = np.searchsorted(classes, c)
        if i >= len(classes) or classes[i] != c:
            raise ValueError("Class label {} not present.".format(c))
        else:
            weight[i] = class_weight[c]
...

在上面的代码片段中，class_weight 被应用于 sample_weight，它被用于一些内部函数，如 _logistic_loss_and_grad , _logistic_loss等:

# Logistic loss is the negative of the log of the logistic function.
out = -np.sum(sample_weight * log_logistic(yz)) + .5 * alpha * np.dot(w, w)
# NOTE: --->  ^^^^^^^^^^^^^^^