python中变量中的函数参数名称

Question

I am using sklearn to train different models. I want to pass the decision tree classifier of sklearn , different values of the same parameter and plot a graph. I want to do this for many such parameters. So, I want to create a general function which can handle all the parameters and their values.

My question is that is there a way to assign the parameter name (not value) to a variable and pass it to my function.

Eg.- Decision tree takes the max_depth , min_samples_leaf etc. arguments. I want to try different values of both parameters one at a time and plot results for both max_depth and min_samples_leaf separately.

Answer 1

Use a dictionary and pass it with ** .

kwargs = {
    "max_depth": value,
    "min_samples_leaf": value,
}
fun(**kwargs)

Answer 2

This solution isn't very "Pythonic", but it's easy to follow. You could just call the function in a loop or nested loop or something similar.

dt = DecisionTreeClassifier(criterion='entropy', min_samples_leaf=150, min_samples_split=100)

Is the standard call to use a decision tree, just loop over the values you want to use and replace min_samples_leaf and min_samples_split

from sklearn.tree import DecisionTreeClassifier
from sklearn.metrics import accuracy_score, roc_curve, auc
from sklearn.model_selection import train_test_split

min_samples_leafs = [50, 100, 150]
min_samples_splits =[50, 100, 150]

for sample_leafs in min_samples_leafs:
    for sample_splits in min_samples_splits:

        dt = DecisionTreeClassifier(criterion='entropy', min_samples_leaf=sample_leafs, min_samples_split=sample_splits)

        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=101)

        dt = dt.fit(X_train, y_train)
        y_pred_train = dt.predict(X_train)
        y_pred_test = dt.predict(X_test)


        print("Training Accuracy: %.5f" %accuracy_score(y_train, y_pred_train))
        print("Test Accuracy: %.5f" %accuracy_score(y_test, y_pred_test))
        print('sample_leafs: ', sample_leafs)
        print('sample_leafs: ', sample_splits)
        print('\n')

Output:

Training Accuracy: 0.96689
Test Accuracy: 0.96348
sample_leafs:  50
sample_leafs:  50


Training Accuracy: 0.96689
Test Accuracy: 0.96348
sample_leafs:  50
sample_leafs:  100


Training Accuracy: 0.96509
Test Accuracy: 0.96293
sample_leafs:  50
sample_leafs:  150


Training Accuracy: 0.96313
Test Accuracy: 0.96256
sample_leafs:  100
sample_leafs:  50


Training Accuracy: 0.96313
Test Accuracy: 0.96256
sample_leafs:  100
sample_leafs:  100


Training Accuracy: 0.96313
Test Accuracy: 0.96256
sample_leafs:  100
sample_leafs:  150


Training Accuracy: 0.96188
Test Accuracy: 0.96037
sample_leafs:  150
sample_leafs:  50


Training Accuracy: 0.96188
Test Accuracy: 0.96037
sample_leafs:  150
sample_leafs:  100


Training Accuracy: 0.96188
Test Accuracy: 0.96037
sample_leafs:  150
sample_leafs:  150

You can make this a function by passing the lists like so

def do_decision_tree_stuff(min_samples_leafs, min_samples_splits):

You call the function like this

 do_decision_tree_stuff([50, 100, 150], [50, 100, 150])