Python DecisionTreeRegressor

Question

I have tried the following code and this error has been occuring to me

Link for DataSet is in link bellow

ValueError ---> line 18 ds1_model.fit(X, y)

ValueError: could not convert string to float: 'Iris-setosa'

  import pandas as pd
from sklearn.metrics import mean_absolute_error
from sklearn.tree import DecisionTreeRegressor
from sklearn.model_selection import train_test_split

url = 'https://raw.githubusercontent.com/jbrownlee/Datasets/master/iris.csv'
ds1 = pd.read_csv(url)
ds1.columns = (['SepalLength' , 'SepalWidth' , 'PetalLength' , 'PetalWidth' , 'ClassLabel'])
ds1_filtered=ds1.dropna(axis=0)

y = ds1_filtered.ClassLabel

ds1_features = ['SepalLength' , 'SepalWidth' , 'PetalLength' , 'PetalWidth']
X = ds1_filtered[ds1_features]

ds1_model = DecisionTreeRegressor()

ds1_model.fit(X, y)

PredictedClassLabel = ds1_model.predict(X)
mean_absolute_error(y, PredictedClassLabel)

train_X, val_X, train_y, val_y = train_test_split(X, y, random_state = 0)
ds1_model = DecisionTreeRegressor()
ds1_model.fit(train_X, train_y)

predicitions = ds1_model.predict(val_X)
print(mean_absolute_error(val_y, predictions))

can you please help to suggest or explain how to fix this?

DataSet Link

Answer 1

As the name ClassLabel implies, the iris dataset is a classification and not a regression one; hence, neither DecisionTreeRegressor is the correct model to use nor mean_absolute_error is the correct metric.

You should use a DecisionTreeClassifier and accuracy_score instead:

from sklearn.datasets import load_iris
from sklearn.tree import DecisionTreeClassifier
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score

iris = load_iris()
clf = DecisionTreeClassifier()

train_X, val_X, train_y, val_y = train_test_split(iris.data, iris.label, random_state = 0)
clf.fit(train_X, train_Y)

pred = clf.predict(val_X)
print(accuracy_score(val_y, pred))

The scikit-learn decision tree classification tutorial using the said dataset can give you more ideas.