尝试在GaussianNB上计算准确性时出现错误

问题描述

我正在尝试获得所创建模型的准确性。我的代码看起来像这样

from sklearn.datasets import fetch_california_housing
from sklearn.model_selection import train_test_split
import numpy as np

data = fetch_california_housing()
c = np.array([1 if y > np.median(data['target']) else 0 for y in data['target']])
X_train,X_test,c_train,c_test = train_test_split(data['data'],c,random_state=0)

gaussian=GaussianNB().fit(X_train,c_train)
pred=gaussian.predict(X_test)    
metrics.accuracy_score(X_test,pred)

此行引发错误metrics.accuracy_score(X_test,pred)

ValueError: Classification metrics can't handle a mix of continuous-multIoUtput and binary targets

搜索解决方案,但找不到任何解决方案。我从其他人那里看到有此问题的一些帖子说,不能使用metric.accuracy ...,因为这是针对分类问题的。但是我的是一个分类问题。

我还尝试了另一种方法pred=score(X_test,pred)发出错误

TypeError: 'numpy.float64' object is not callable

感谢您的帮助

---------------------更新-------------

X_test

[[   4.1518       22.            5.66307278 ...    4.18059299
    32.58       -117.05      ]
 [   5.7796       32.            6.10722611 ...    3.02097902
    33.92       -117.97      ]
 [   4.3487       29.            5.93071161 ...    2.91011236
    38.65       -121.84      ]
 ...
 [   3.6296       16.            3.61684211 ...    1.88631579
    34.2        -118.61      ]
 [   5.5133       37.            4.59322034 ...    3.00847458
    33.9        -118.34      ]
 [   4.7639       36.            5.26181818 ...    2.90545455
    37.66       -122.44      ]]

Pred

[1 1 1 ... 1 1 1]

解决方法

似乎可以使用c_test:

accuracy_score(c_test,pred) # 0.743798

另一种方法是:

1 - ((c_test != pred).sum() / X_test.shape[0]) # 0.743798