December 2018
Intermediate to advanced
318 pages
8h 28m
English
This approach uses a model selection algorithm to find the best performing model for each use case. The bucket of models needs to be tested across many use cases to derive the best model by weighting and averaging. Similar to the BMC method, they choose models in the bucket by methods of cross validation. If the number of use cases are very high, the model that takes more time to train should not be taken from the selection. This selection of considering fast-learning models is also known as landmark learning.
Read now
Unlock full access