Sai Sindhura Tipirneni
Reputation: 121
sklearn KNN regression does not give 0 training error for k=1
I am trying out sklearn's KNN regressor. I expect it to have 0 training error for k=1. I am getting a mean square training error of 0.071. Is this possible?
Upvotes: 0
Views: 186
Answers (1)
Sai Sindhura Tipirneni
Reputation: 121
Figured out. When two training samples have same input and different output, the regressor doesn't consider both for k=1. It considers only one of them.
Upvotes: 1
Related Questions
- Fit function in sklearn KNN model does not work: 'n_neighbors does not take <class 'float'> value, enter integer value'
- My own KNN function doesn't give same result as scikit-learn
- Support vector regression from sklearn gives flat prediction
- K-NN: training MSE with K=1 not equal to 0
- KNN model, accuracy(clf.score) returns 0
- How to resolve Sklearn KNN Not fitted Error
- my data has 14 attributes and 303 observations but when applying knn value of k greater than 1 is giving error
- What is the Kernel for the KNN regressor in sklearn python?
- Error when using KNeighborsClassifier using sklearn
- KNN only predicting one class. Using python and sklearn