Fit logistic regression and return top features
Company: Roblox
Role: Data Scientist
Category: Machine Learning
Difficulty: hard
Interview Round: Online Assessment
You are given:
- `X`: a 2D numeric array where **each row is a feature** and each column is an observation (shape: `n_features x n_samples`).
- `feature_names`: a list of length `n_features`.
- `y`: a binary outcome array of length `n_samples` with values in {0, 1}.
Task:
1) Fit a (binary) **logistic regression** model to predict `y` from `X`.
- Include an intercept.
- No regularization unless explicitly stated.
2) Rank features by the absolute value of their fitted coefficients (exclude the intercept).
3) Return the **top 3 feature names** by `|coef|` (break ties by lexicographic order of `feature_names`).
Output:
- A list of 3 strings (feature names).
Overview: This question evaluates understanding of binary logistic regression, coefficient-based feature importance, and practical model-fitting considerations such as intercept inclusion and input shape handling.
Read the full Roblox Data Scientist interview experience this question came from