Making the Most of Machine Learning End-to-End
In this session we go end-to-end to construct a machine learning web service using Microsoft Azure Machine Learning (ML). The session begins by reading in a data set from Azure blob storage and a second data set from a public website. Together we use Azure ML to clean the data set, explore the data, create and select features in the data set, then evaluate several different machine learning algorithms to select the optimal model to put into production as a machine learning web service. Along the way we discuss features of Azure ML, such as imputing missing values in the data set, running arbitrary R script to compute statistics and create features, to get a broad view of the functionality available to a data scientist using Azure ML.