Tuning Your Models with AMLS

Tuning your models is an important step in your data science journey. The objective of a data science workload is to provide the best model on unseen data in the shortest duration of time. In order to provide a reliable model, not only are you required to tune the features that are the inputs to your model but you also need to tune the parameters of your model itself. Model parameters, also known as hyperparameters, can have a significant impact on the performance of your trained model. Tuning a model can take a lot of effort and involves trial and error. Several frameworks can be leveraged to automate this task. AMLS provides this functionality, which we will explore in this chapter. AMLS allows you to define model parameters that should be tuned to find the best model through the use of a special type of job referred to as a sweep job. These hyperparameters will be defined for a given AMLS job, and AMLS will run many trials and determine the best model...