View the tuning results

YukiN

I would like to understand Optimization strategies.

SESSION 1 is the result of setting multiple hyperparameters (default: 8 values), and SESSION 2 is the result of setting only one of those parameters, but SESSION 2 has a higher score.

I thought that when I created several models with hyperparameters, the one with the best score specified in Metrics would be selected, am I right?

 

Alexandru

You understanding is correct :

https://doc.dataiku.com/dss/latest/machine-learning/advanced-optimization.html#visualization-of-hyperparameter-search-results