Category : supervised-learning

I developed a model in Python, using sklearn, to apply machine learning algorithms with a classified dataset. I applied, for example, the RandomForestClassifier algorithm. I got the values for accuracy, precision, recall, among others. Now, I want to put this model into production. I want to send a record to the model and receive the ..

Read more

I know that scikit-learn provides for roc_auc_score function to calculate roc-auc score metric, including for multiclass classification. However, this function can only be used with classifiers that support the predict_proba probability estimate methods such as decision trees. When using a classifier that does not provide that method, such as Perceptron for example, one has to ..

Read more

Hi I have a base with p=40 and n=11750, I want to build a model with Ridge classifier (because the objective is predictive if one person is going to be defaulter=0 o not defaulter=1) that select the 10 most representative features in a 10-fold Cross Validation and then calculate de accuracy of the model. For ..

Read more

I am trying to visualize the decision boundary for a classification task. While i’ve read some good paper here (Plot k-Nearest-Neighbor graph with 8 features?) and here (https://ogrisel.github.io/scikit-learn.org/sklearn-tutorial/auto_examples/tutorial/plot_knn_iris.html) i encounter an error: MemoryError: Unable to allocate 737. GiB for an array with shape (98903372450,) and data type float64 Can i reduce my sample or sth ..

Read more