Posted On: Apr 5, 2016
You now have the ability to shuffle your training data with HAQM Machine Learning (HAQM ML) for improved ML model quality. HAQM ML creates better ML models when input rows are presented to the ML algorithm in a random order. With this new training parameter option, you can have HAQM ML shuffle the order of the rows of your training data. Data shuffling will be turned on by default in the HAQM ML console when you create an ML model. You can control this option through the console, or turn data shuffling on through the API and SDK.
For more information on data shuffling, visit the HAQM ML documentation.