Explain bias-variance and evaluate a classifier
Company: Microsoft
Role: Machine Learning Engineer
Category: Machine Learning
Difficulty: easy
Interview Round: Technical Screen
You are interviewing for an Applied Scientist internship. Answer the following ML foundations questions.
## 1) Bias–variance
- Define **bias** and **variance** in supervised learning.
- Explain the **bias–variance tradeoff** and how it relates to **underfitting vs. overfitting**.
- Give 2–3 practical ways to reduce:
- high bias
- high variance
## 2) Classification metrics
- Define **accuracy, precision, recall, F1**.
- Explain when accuracy is misleading.
- Given a confusion matrix (TP, FP, TN, FN), show how you would compute the metrics and choose which one to optimize for an imbalanced problem.
## 3) Confidence intervals
- What is a **confidence interval (CI)**?
- Suppose you evaluated a binary classifier on a test set of size \(n\) and observed accuracy \(\hat{p}\). Describe how you would compute a **95% CI** for the true accuracy and what assumptions are required.
- Name at least one alternative method to build a CI if assumptions are weak (e.g., small sample size or correlated examples).
Quick Answer: The question evaluates foundational machine learning competencies including the bias–variance tradeoff, model underfitting versus overfitting, classification metrics (accuracy, precision, recall, F1), and statistical evaluation of classifiers via confidence intervals.