machine-learning - Scikit-learn - KNeighborsClassifier 的用户定义权重函数

Question

我有一个 KNeighborsClassifier，它根据 4 个属性对数据进行分类。我想手动对这 4 个属性进行加权，但总是遇到“操作数无法与形状 (1,5) (4) 一起广播”的情况。

关于 weights : [callable] : a user-defined function which accepts an array of distances, and returns an array of the same shape containing the weights. 的文档很少(来自here)

这就是我现在所拥有的:

    for v in result:
        params = [v['a_one'], v['a_two'], v['a_three'], v['a_four']]
        self.training_data['data'].append(params)
        self.training_data['target'].append(v['answer'])

    def get_weights(array_weights):
        return [1,1,2,1]

    classifier = neighbors.KNeighborsClassifier(weights=get_weights)

Answer 1

可调用 sklearn 权重的说明

import numpy as np
from sklearn.neighbors import KNeighborsClassifier

创建模型训练的样本数据

df = pd.DataFrame({'feature1':[1,3,3,4,5], 'response':[1,1,1,2,2]})

y = df.response
# [1,1,1,2,2]

X_train = df[['feature1']]
# [1,3,3,4,5]

定义自定义距离函数(打印输入数据结构)

def my_distance(weights):
    print(weights)
    return weights

将 my_distance 中传递的模型定义为权重的可调用

knn = KNeighborsClassifier(n_neighbors=3, weights=my_distance)

knn.fit(X_train,y)

knn.predict([[1]])
# [[ 0.  2.  2.]]
# array([1])

说明:显示与预测值1最接近的3个邻居(n_neighbors=3)

X_train 中距离 1 最近的三个邻居:

1, 3, 3 

距离:

[[ 0.  2.  2.]]

1 - 1 = 0 
3 - 1 = 2
3 - 1 = 2

预测类别:

array([1])