The goal of Analytics is to support better business decisions through the generation of relevant and accurate inferences made accessible to the right people in an understandable and timely manner. From the standpoint of Analytics, the promise of machine learning (ML) is the promise of more. More relevant information, more accurate inferences, more understandable relations and more timely delivery – all gleaned through deeper learning.
But there is a catch. Though ML (e.g., deep neural nets) can arguably learn better than native Analytics functions (e.g., least squares), ML realizes its ‘magic’ through specialized offline training exercises that are highly sensitive to the specifics of the training data. Integrating the results of ML-based advanced analytics (e.g., from text) into production Analytics and ensuring that the ML training processes are appropriately guided by knowledge that already exists and is constantly changing, is a challenge that has kept most organizations from realizing the full benefits that ML has to offer.
In this session you will learn about:
- The major challenges for Analytics to successfully leverage ML
- How to choose where to start
- How to use ML to extract relevant information from text that can be integrated into other analytic processes