Effective Amazon Machine Learning

Conclusion

Cross-validation may be too time-consuming to implement via shell only. There are many files to create and coordinate. There are simpler ways to implement cross-validation with libraries such as scikit-learn for Python or Caret for R, where the whole model training and evaluation loop over several training and validation sets only requires a few lines of code. However, we showed that it was possible to implement cross-validation with Amazon ML. Cross-validation is a key component of the data-science workflow. Not being able to do cross validation with Amazon ML would have been a significant flaw in the service. In the end, the AWS CLI for machine learning is a very powerful and useful tool to conduct sequences of trials and compare ...