Hyperopt is not suitable for tuning hyperparameters in time-series forecasting models.

The challenges include handling temporal dependencies and ensuring that the search space accounts for time-related features. Hyperopt can be adapted by defining a search space that includes parameters relevant to time-series analysis and using techniques like rolling window validation to manage temporal dependencies.

The only challenge is to increase the number of trials to ensure comprehensive coverage of the hyperparameter space.

Time-series forecasting models do not require hyperparameter tuning, so Hyperopt should not be used.