The most common recommendation engines we interact with daily involve social media (the people we may know), retail (the products we may like), and entertainment (the music, clips and/or movies presented next in our streams). Yet these are just the tip of the recommendation iceberg. When we look deeper into the inner workings of the largest organizations, enterprises and agencies in the world with critical business and mission decisions to make – based on a constant flow of ever changing, highly inter-connected data – we find much more complex and significant types of recommendations coming into play through prescriptive recommendation engines.
The ability to quickly and confidently make informed decisions is paramount in rapidly evolving mission scenarios. Designing and building reliable prescriptive recommendation engines starts with taking advantage of real-time connected data; the fuel for what will drive forward-thinking and innovative mission goals within an organization. ONgDB is a free and open source native graph database that is optimized for working with highly connected data and excels as a backing store for recommender systems.
A key component in prescriptive recommendation engines is the confidence, governance and justification providing the audit trail for why a recommendation was made. Without the decision path, the human in the flow cannot quickly and confidently act on the recommendation.
As important as it is to arrive at a decision with a high level of confidence, providing a feedback loop to integrate the actual outcome from the is equally important. Feeding the outcomes from a decision back into the prescriptive recommendation framework enables fully automated actions to be taken when confidence, governance and justification metrics reach necessary thresholds.
Whether it involves
- leveraging direct or indirect mission intelligence,
- insights through social media platforms,
- analysis of business supply chain details from the manufacturing plant to the logistics network,
- or inferring relationships according to an activity and utilizing the network to determine the confidence in that assertion
GraphGrid’s Connected Data Architecture improves mission outcomes through a prescriptive recommendation framework that provides authorized resources critical analysis and prioritization of critical scenarios to assess and advised decisions to make.
For example, if you are intent of finding insurance, medical, or financial fraud, there are a number of understood graph patterns associated with these fraud scenarios that surface within the data. These patterns can be used to proactively uncover transactions and rank them by priority for immediate examination of the most critical ones. Such detection is essential for analysts because it orders likely fraud events by a confidence score, saving time of a high value resource by examining transactions that do not meet specific standards.
GraphGrid Connected Data Platform is essential for enabling real-time prescriptive recommendations based on the real-time awareness and understanding of your mission derived from your complex data flows. Our expert connected data architects at GraphGrid can work with you to design and optimize your data – fed into our real-time prescriptive recommendation engine framework – producing greatly improved mission outcomes.