Benefits of Hybrid Analytica

For example, in control systems where it is important to capture non-linearities to avoid excessive model complexity and overfitting, hybrid modeling can be useful. In this case the analytical model is used to provide some of the error structure and the remaining error is then learned data driven. This can reduce the training data demand and enable a less complex and more interpretable error model than the pure analytical model.

Another use is in ML-in-the-loop, where machine learning methods are applied within the control loop of industrial systems. This can help improve estimation of the system state and performance by leveraging information in the data as well as traditional domain knowledge and control theory. This can yield more robust and efficient controllers than a purely ML approach.

The fourth use is in fusion modeling, which is the combining of at least two different types of data to train an ML model. For example, in process industry control systems, the combination of process data with sensor and actuator measurements can allow for a more accurate model. This can result in more efficient operation and reduced maintenance costs.

A further benefit of hybrid analytics is that it allows users to leverage business intelligence assets across multiple platforms. In this way the organization can deliver a more consistent and user friendly interface without having to build a new solution for every new type of visualization or reporting. This is a major benefit for organizations with existing systems that have been in use for some time, and for whom it would be cost prohibitive to rebuild their business analytics interface.

An example is a portal in which reports and visuals that are part of daily workflows but stored in different business intelligence platforms can be published to a web based interface. This would allow for the creation of a single listing of hyperlinks under logical groupings like sales or finance, which allows end users to easily access the data and visualizations they need from multiple systems in one location without having to remember where each report or visual is located within the various individual systems.

Hybrid Hybrid Analytica also provides a more unified view of cross-environment business analytics through durable object storage. This can allow for a consolidated analytics platform that can be deployed on-prem or in the cloud, with the ability to scale out and back as required by business needs. This can be particularly beneficial in environments where the system architecture is evolving rapidly to include hybrid components. This is currently common in the oil and gas industries where companies need to be able to run business intelligence and machine learning applications on premises as well as within their cloud based platforms. This requires a flexible, scalable and secure environment. This is what we are delivering through our Hybrid Analytica.