THE TECHNOLOGY BEHIND HUGIN PREDICTIT

HUGIN PredictIT is a multitenant software stack designed as a Software-as-a-Service solution. PredictIT can be installed and run locally on the company’s own systems or hosted externally if preferred.

The solution uses Bayesian networks (an advanced statistical model) to learn statistical relationships between sensor measurements based on a combination of data and knowledge. This model is used to identify situations that deviate from normal operation, that can be an early indication that a possible breakdown is on the way. The method can identify which sensor measurements that creates a possible error. This information can be used in connection with troubleshooting if an error occurs.

The uniqueness of Bayesian networks is that they allow combining data and knowledge to build the best model. It is therefore possible to start using PredictIT without a single measurement. The implementation of HUGIN PredictIT does not necessarily start with a major data collection.

It is possible to use PredictIT on single machines, production lines and factory level.

HUGIN’s general software package is used to solve very different problems in several areas. Read more about HUGIN software here.