ProMoS NG - Building Intelligence

Over 30 different algorithms are implemented in ProMoS NG. The system evaluates the suitable algorithms completely automatically.

 

Functions

Usable functions

The BI module of ProMoS NG opens up a wide range of possible applications for machine learning:

  • Create tasks: Users can define explanatory variables (e.g. process data points), temporal resolutions and time spans. A variable limit value is calculated on this basis, which depends on both the specified variables and the temporal position of the values. The ProMoS-BI module independently evaluates the optimal models to ensure the best possible results.
  • Another option is a specially set up test environment: the user can make all the settings here independently, select suitable ML models and have the calculations carried out.
  • Productive evaluations: The calculations are not just carried out once, but in regular cycles. This enables continuous monitoring of the variables.
     

Models

Machine learning models

The following ML algorithms are currently implemented:

  • AdaBoostRegressor
  • BaggingRegressor
  • DecisionTreeRegressor
  • ExtraTreesRegressor
  • GammaRegressor
  • GaussianPricessRegressor
  • GradientBoostingRegressor
  • HistGradientBoostingRegressor
  • HuberRegressor
  • KNeighborsRegressor
  • KernelRigde
  • Lars
  • LarsCV
  • LassoCV
  • LassoLarsCV
  • LassoLarcIV
  • LinearRegression
  • LinearSVR
  • MLPRegressor
  • NUSVR
  • OrthoganaMatchingPursuit
  • OrthoganaMatchingPursuitCV
  • PiecewiseRegressor
  • RANSACRegressor
  • RadiusNeighborsRegressor
  • RandomForestRegressor
  • RidgeCV
  • SVR
  • TheilSenRegressor
  • XGBoos