I came across this question during an assessment:
A telecommunications company predicts customer churn based on usage patterns, customer demographics, and customer service interactions. However, the company suspects some input variables may have outliers that could influence the model's performance.
Which technique can help mitigate the influence of outliers in multiple linear regression?
From what I can remember, the options were
Elastic Net Regression
Isolation forest?
Option
Option
I chose elastic net as answer but it was marked incorrect. ChatGPT and Gemini chose elastic net as well.
What is the correct answer and why?