Predictive Modeling Flashcards
7 cards from real MAC practice questions. Tap to flip, then mark Knew It or Still Learning — missed cards come back until you master them.
Read the first 7 Predictive Modeling flashcards as text
What does a lift curve measure in the context of a marketing predictive model?
Answer: How much better the model performs compared to random targeting when ranking prospects
A lift curve shows how much more likely model-selected prospects are to respond compared to a random selection, helping prioritize outreach.
In attribution modeling, the 'time decay' model assigns credit by:
Answer: Giving more credit to touchpoints closer in time to the conversion
The time decay attribution model gives more credit to touchpoints that occurred closer to the conversion event, reflecting their greater influence.
Which evaluation metric is most informative when the cost of false positives and false negatives are very different?
Answer: F1 Score
F1 Score balances precision and recall, making it more informative than accuracy when misclassification costs are asymmetric.
A company segments customers using K-Means clustering before building a predictive model. What is the primary benefit of this approach?
Answer: It allows the team to build separate, more tailored models for each customer segment
Segmenting customers before modeling allows separate, specialized models per cluster, capturing segment-specific behavior better than a single global model.
What is 'data leakage' in predictive modeling?
Answer: Including information in the model that would not be available at prediction time
Data leakage occurs when future or target-derived information is included in model training, causing artificially inflated performance that won't hold in production.
Which hyperparameter controls the number of trees in a Random Forest model?
Answer: n_estimators
The n_estimators hyperparameter specifies how many decision trees are built and aggregated in a Random Forest ensemble.
A predictive lead scoring model outputs a probability of 0.82 for a prospect. What does this mean?
Answer: There is an 82% estimated probability that this prospect will convert
A probability output of 0.82 means the model estimates an 82% likelihood that this specific prospect will convert, based on their feature profile.