在这个循环中如何计算 AUC、Precision、Recall

扎眼的阳光 2年前 python 203

原文标题 ：How is AUC, Precision, Recall being calculated in this loop

此代码块创建一个模型字典，并通过评分和模型迭代创建三个键值对。

{0},{1}标识的格式由eval()方法使用j（分数值）和index匹配的model_name字符串映射到model名称。

{0}被j替换，{1}被model_name[model.index(i)]替换。

只有字符串中的变量，即{0}和{1}被评估，并且不是变量或运算符的无关字符，即Mean_和_CV被忽略。那么实际的AUC，Precision和Recall值是如何计算的？

model = ['Logistic Regression','KNN','Gaussian NB','Decision Trees','Random Forest','Ensemble']
scoring = ['AUC','Precision','Recall']
model_name = ['Logit','KNN','NB','tree','forest','ensemble']
model_list = []

for i in model:
    for j in scoring: 
        model_dic = {'Model': i,'Scoring':j, 'Score':eval('Mean_{0}_{1}_CV'.format(j,model_name[model.index(i)]))}
        model_list.append(model_dic)

输出

# Model | Scoring | Score
# Logistic Regression   AUC 0.957516
# ...

原文链接：https://stackoverflow.com//questions/71463514/how-is-auc-precision-recall-being-calculated-in-this-loop

我来回复

Code-Apprentice 评论

该回答已被采纳！

model_dic = {'Model': i,'Scoring':j, 'Score':eval('Mean_{0}_{1}_CV'.format(j,model_name[model.index(i)]))}

这条线一次做的太多了。要了解它在做什么，请将其分解为更小的部分：

formatted = 'Mean_{0}_{1}_CV'.format(j,model_name[model.index(i)])
evaluated = eval(formatted)
model_dic = {'Model': i,'Scoring':j, 'Score':evaluated}

现在您可以添加print()语句以更好地了解正在发生的事情：

formatted = 'Mean_{0}_{1}_CV'.format(j,model_name[model.index(i)])
print(formatted)
evaluated = eval(formatted)
print(evaluated)
model_dic = {'Model': i,'Scoring':j, 'Score':evaluated}
print(model_dic)

我建议您阅读format()和eval()函数以了解输出。

2年前 0条评论