Walmart·Data Scientist·Technical Phone Screen
- How would you build a model to predict ad conversion rate for the next time period using historical campaign data like impressions, clicks, and conversions?
- Should the conversion rate target variable be transformed before modeling, and if so, why?
- If you had access to more data, what additional features would you incorporate into the model?
- How would you evaluate the performance of this conversion rate prediction model?
- Walk me through the loss function used in logistic regression.
- Explain PCA in terms of eigenvalues and eigenvectors, and describe the assumptions the method relies on.
“I talked through a few modeling options and landed on using lagged features from prior periods.”