Azure ML Workbench

Source: Microsoft

Azure Machine Learning (ML) Workbench is a new tool (in preview) used to work with the Azure ML service to build machine learning models, which use existing data to predict future results. Workbench includes several features to simplify the preparation of data to train machine learning models, including Programming Synthesis using Examples (PROSE), which derives formatting rules from desired outcome examples specified by the user.

Shown here, existing data records include the Call Time column, which is expressed as the time of day. The developer is building a model that requires a column specifying the half-hour time window within which the Call Time occurs. The developer types the example “5:00PM-5:30PM” in the first row, and Workbench derives how that text is produced from the data in the Call Time column and fills in the corresponding time windows for the remaining data records. The developer could provide examples in additional rows to tune the results if necessary.