在运行随机森林回归模型时,我在运行混淆矩阵时遇到了错误。虽然模型能够预测y_pred,但我无法检查准确性,因为我无法运行准确性矩阵。
数据类型:
type(y_test)Out[124]: numpy.ndarraytype(y_pred)Out[125]: numpy.ndarray
以下是我用于混淆矩阵的代码:
from sklearn.metrics import confusion_matrixmatrix = confusion_matrix(y_test, y_pred)print(matrix)from sklearn.metrics import accuracy_scoreaccuracy = accuracy_score(y_test, y_pred)
错误:
confusion_matrix(y_test, y_pred)Traceback (most recent call last):File "<ipython-input-121-7a3dbcfd18a0>", line 1, in <module>confusion_matrix(y_test, y_pred)File "/Users/srikantswamy/anaconda3/lib/python3.6/site-packages/sklearn /metrics/classification.py", line 253, in confusion_matrixy_type, y_true, y_pred = _check_targets(y_true, y_pred)File "/Users/srikantswamy/anaconda3/lib/python3.6/site-packages/sklearn /metrics/classification.py", line 88, in _check_targetsraise ValueError("{0} is not supported".format(y_type))ValueError: continuous is not supported
回答:
简单回答:混淆矩阵通常用于描述分类模型(或“分类器”)的性能。在您的案例中,您尝试为连续变量查找混淆矩阵。对于回归问题,应使用均方误差。请查看此讲义PDF:http://www.it.uu.se/edu/course/homepage/sml/lectures/lecture3_handout.pdf