Databricks Machine Learning Associate Question 200
Select 3You are training a machine learning model in Databricks and want to tune its hyperparameters using random search. Which of the following steps are required to implement random search in this scenario?
- A
Define a search space for the hyperparameters to explore
- B
Manually select a subset of hyperparameter combinations to evaluate
- C
Use a random sampling strategy to select hyperparameter combinations from the search space
- D
Train the model for every possible hyperparameter combination in the search space
- E
Evaluate the model performance for each selected hyperparameter combination
Show answer and explanation
Correct answers: A, C, E
Explanation
Random search is a hyperparameter tuning method where hyperparameter combinations are selected randomly from a predefined search space. The key steps are defining the search space, using a random sampling strategy to pick combinations, and evaluating the model performance for each selected combination. Random search is computationally more efficient than exhaustive grid search because it does not evaluate all possible combinations in the search space.
- A. Correct.
Correct: Defining a search space is essential for random search because it specifies the range or distribution for each hyperparameter to sample from.
- B. Incorrect.
Incorrect: Random search automates the selection of hyperparameter combinations. You do not manually select them.
- C. Correct.
Correct: Random search involves using a random sampling strategy to choose combinations of hyperparameters from the defined search space.
- D. Incorrect.
Incorrect: Training the model for every possible combination is exhaustive search, not random search. Random search only evaluates a subset of combinations.
- E. Correct.
Correct: Evaluating the model performance for each sampled hyperparameter combination is a necessary step to determine the best-performing configuration.