Tackling the Problem of Telematics, Big Data and Field Experience from Multiple Directions
The Houston Business Journal recently wrote about the company Verdande Technology and their use of case-based reasoning (CBR) for health care and oil/ gas exploration. It’s a sophisticated system that takes a different approach to what is essentially a big data problem to help improve medical and drilling decisions with real-time predictive analytics.
Both Verdande and CaseBank use CBR as an enabling technology to leverage, and learn from, previous experience to deliver fast, accurate insight to difficult problems—ultimately improving the efficiency and consistency of customer/product support. Comparing the use of CBR by Verdande and CaseBank reveals the many ways this technology can be applied.
For oil exploration, Verdande uses CBR in the operations centre to run pattern matching over real-time data streams from operating equipment. It thrives on LOTS of data, and the pattern matching runs continuously. Data is matched against known problems and the relative risks of each are computed and displayed. Decision makers can monitor the risk levels and take action whenever necessary to keep risks within manageable ranges. When a risk threshold is crossed, alerts are triggered and operators are notified that urgent action is required.
CaseBank, on the other hand, is much less dependent on large amounts of data; rather, it’s used by front line technicians or call center personnel to improve troubleshooting and repairs of complex equipment. CaseBank uses CBR to consider all known root causes for each equipment failure, and establishes an optimal fault isolation sequence that continuously changes as new information or feedback becomes available.
Both Verdande and CaseBank provide just-in-time decision support to improve the efficiency and consistency of equipment operations and repairs—maximizing uptime. While Verdande strives to keep problems from occurring, CaseBank ensures a fast resolution when they do occur. While Verdande warns when failures are imminent, CaseBank explains what, why and how the failure occurred—ensuring problems are verified, prioritized and quickly resolved.
Like two key ingredients of a perfect recipe (chocolate and peanut butter?), these are two great tools that could work great together. Imagine the potential that a combined system could offer:
- Verdande identifies a problem with low but increasing risk and hands it to CaseBank. CaseBank suggests a plan to investigate and confirm the impending problem. Given the low urgency, as estimated by Verdande, the plan maximizes accuracy and cost-effectivity.
- Verdande identifies the same problem, but with greater risk. CaseBank creates a high-urgency plan that emphasizes accuracy and rapid response.
- An inevitable failure occurs. Verdande’s hypotheses provide additional clues (input symptoms) to guide CaseBank’s troubleshooting process, speeding return to service.
I’d love to put a system like this into production. Any takers?
For more about CaseBank’s use of case based reasoning (CBR) please click here.