Example Applications

Relational databases are the defacto standard for storing and querying structured data that are critical to the operation of most businesses. They capture a huge wealth of information that can be used for data-driven decision making, and for identifying risks and opportunities. Example applications include recommendation, CRT prediction, healthcare analytics, and etc, which are highlighted as follows:

• Frappe - App Recommendation
• MovieLens - Movie Recommendation
• Avazu - Click-Through Rate Prediction
• Criteo - Display Advertising Challenge
• Diabetes130 - Diabetes Readmission Prediction

Besides providing more accurate prediction results, ARMOR also supports both global and local interpretability, which correspond to understanding the whole logic of the learning model and providing reasons for making a specific decision, respectively. Let us take healthcare analytics for example (see the above figure), ARMOR can identify important features globally for the whole model and support local feature attribution on a per-patient basis, which help engender more personalized analytics and healthcare management.