Problem and Data

Five questions. Choose one answer for each, then submit to see your score and explanations.

The practice target is 80%. You can retry as often as you like, and every lesson stays accessible.

You can practice as a guest. Sign in before starting an attempt to save your scores.

1. If the dispatcher predicts whether a delivery will exceed a ten-minute delay limit, what type of prediction is the output?A. ClassificationB. Image segmentationC. Clustering without a targetD. A physical camera calibration

2. What should validation records be used for?A. To fit all model parameters and also claim a final unbiased scoreB. To compare candidates and choose settings before using the held-aside test recordsC. To reveal test answers before trainingD. To replace the training records entirely

3. When should feature preprocessing be fitted to avoid leaking evaluation information?A. Once on the full dataset before splittingB. Separately on the test labelsC. On training data within each split or training fold, then applied to its held-out dataD. Only after the final test predictions are announced

4. A thousand chronological deliveries are divided into 700 training and 150 validation records. How many remain for the held-aside test period?A. 50B. 100C. 150D. 300

5. A prediction is required before departure. Which proposed input causes information leakage?A. Planned routeB. Forecast available before departureC. Scheduled departure timeD. Actual arrival time recorded after the delivery

Review this module from its first lesson →

Problem and Data

Your result