Databricks Machine Learning Associate Question 610
Single answerYou are building a machine learning model to rank products on an e-commerce website based on their relevance to user queries. The goal is to optimize the ranking order so that the most relevant products appear at the top of the search results. Which evaluation metric is the most appropriate to assess the performance of your model?
- A
Mean Squared Error (MSE)
- B
F1 Score
- C
Mean Reciprocal Rank (MRR)
- D
Area Under the Receiver Operating Characteristic Curve (AUC-ROC)
Show answer and explanation
Correct answer: C
Explanation
The objective is to optimize the ranking of products based on their relevance to user queries. Mean Reciprocal Rank (MRR) is the most suitable metric as it specifically evaluates the quality of ranking by focusing on the position of the most relevant item in the ranked list. Other metrics like MSE, F1 Score, and AUC-ROC are designed for regression or classification tasks and do not address the ranking problem directly.
- A. Incorrect.
Mean Squared Error (MSE) is used for regression tasks and measures the average squared difference between the predicted and true values. It is not suitable for ranking tasks.
- B. Incorrect.
F1 Score is used for classification tasks to measure the balance between precision and recall. It does not evaluate the quality of ranking.
- C. Correct.
Mean Reciprocal Rank (MRR) is specifically designed for ranking systems. It evaluates how well the model ranks the most relevant items at the top of the list, making it the most appropriate metric in this scenario.
- D. Incorrect.
Area Under the Receiver Operating Characteristic Curve (AUC-ROC) is used to evaluate classification performance, particularly the trade-off between true positive and false positive rates. It is not designed for ranking tasks.