January 2020
Intermediate to advanced
346 pages
9h 8m
English
The conventional, exhaustive grid search, covering all 200 possible combinations, needs to be run next. The search results indicate that the best combination was n_estimators = 70, learning_rate ≈ 0.359, and algorithm = 'SAMME.R'.
The classification accuracy that we achieved with these values is about 93%, which is a vast improvement over the original 65%. The search runtime was about 32 seconds:
performing grid search...best parameters: {'algorithm': 'SAMME.R', 'learning_rate': 0.3593813663804626, 'n_estimators': 70}best score: 0.9325842696629213Time Elapsed = 32.180874824523926
Next comes the genetic-powered grid search. Will it match these results? Let's find out.
Read now
Unlock full access