In this notebook, We will evaluate the five best known machine learning algorithms to determine which one is the best fit for this given telecom churn binary classification.
-
In the first phase of this notebook, we explore and analyze the dataset and acquire a basic understanding of it.
-
In the second phase, we preprocess the dataset and split into train and validation.
-
In the third phase, we will assess various classifiers, including k-nearest neighbor, logistic regression, support vector machines, random forest classifiers, and naive bayes classifier and also well as we analyze the confusion matrix of those as well to determine which clasifire produce the best results.
-
In final phase, we will be compare all the models performance using AUC score and ROC curves to get the more visual comparison.