ABSTRACT
The two most dominant forces in technology in the 21st century are apps and artificial intelligence (AI). We would like to use analytical methods such as AI, machine learning, and predictive analytics to improve business outcomes and user experience in real time. But integrating analytical methods into apps is difficult. There are many roles involved, including developers, data scientists, and business stakeholders. And apps are almost always event-driven, so it may be difficult for a data scientist to understand how to apply analytical models. We have implemented a previously proposed framework that instruments the app with decision points that capture data and control the application. The framework enables the running of experiments to provide statistically valid training data, which can be used to train models that can then be linked back to the decision points to control the app to improve behavior. We show that the framework is feasible and provides an easy way to gain an understanding of the application and its data and to generate training data. The framework provides crucial knowledge, including the decisions made in the application, the data available in real time, and the process involved in incorporating intelligence into the application. Oracle and Java are registered trademarks of Oracle and/or its affiliates. Other names may be trademarks of their respective owners.
