Databricks Machine Learning Associate Question 546
Single answerYou are working on a machine learning project to predict whether a customer will churn (binary classification). The dataset has approximately 10,000 rows and contains both numerical and categorical features. After initial analysis, you notice class imbalance in the target variable, with 90% of the customers classified as 'not churned' and 10% classified as 'churned'. Which of the following algorithms would be the most appropriate to start with in this scenario?
- A
Logistic Regression
- B
K-Means Clustering
- C
Random Forest
- D
Linear Regression
Show answer and explanation
Correct answer: C
Explanation
Random Forest is the most appropriate choice in this scenario because it is a powerful supervised learning algorithm well-suited for classification tasks, including those with imbalanced datasets. While Logistic Regression could also be used, additional preprocessing steps would be needed to handle the class imbalance effectively. K-Means Clustering and Linear Regression are not applicable for this binary classification problem.
- A. Incorrect.
Logistic Regression is well-suited for binary classification problems but may not perform well with imbalanced datasets unless proper techniques like class weighting or resampling are applied.
- B. Incorrect.
K-Means Clustering is an unsupervised algorithm used for clustering tasks, not for supervised classification problems like churn prediction.
- C. Correct.
Random Forest is a robust algorithm for classification tasks. It handles imbalanced datasets well, especially when combined with techniques like class weighting or sampling. It can also manage both numerical and categorical features effectively.
- D. Incorrect.
Linear Regression is designed for regression tasks and is not suitable for binary classification problems like this one.