Handle Missing Values and Outliers in Machine Learning
Overview
This question evaluates skills in data preprocessing, robust model development, evaluation metric selection, algorithmic understanding, and hyperparameter tuning—covering handling missing values and outliers, choosing classification and regression metrics, explaining an ML algorithm's mechanics, and describing key XGBoost parameters in the Machine Learning domain. It is commonly asked to assess reasoning about real-world tabular data and production-ready trade-offs, and it tests both conceptual understanding and practical application by probing data treatment decisions, metric implications, algorithm behavior, and parameter effects.
Technical Screening: Model Development Discussion
Context
You are building classification and regression models on tabular business data with missing values and potential outliers. You must choose appropriate data treatments, evaluation metrics, and modeling approaches suitable for production.
Tasks
-
Missing Values
- Describe at least two methods to handle missing values in a training set.
- For each method, state the pros and cons.
-
Outliers
- Provide two strategies for treating outliers.
- Explain when you would prefer each strategy.
-
Model Evaluation
- Which metrics would you use to evaluate classification and regression models?
- Justify your choices and note any pitfalls.
-
Algorithm Walkthrough
- Pick one machine-learning algorithm and explain how it works step by step.
-
XGBoost Hyperparameters
- List the key hyperparameters in XGBoost and explain their impact on the model.
Solution
Login required