GPC PVA

OPTIMAL SELECTION OF RELEVANT VARIABLES

GPC PVA (Principal Variables Analysis) is a data analysis software particularly usefull in industrial production to: 

REDUCE THE PRODUCT VARIABILITY:

  • •To research the major causes of product variability coming from the process.
  • •To identify the product variabilities independent from available process measures.
  • •To identify the process measures where the variability does not poorly influence the product characteristics (robustness).

REDUCE THE PROCESS VARIABILITY:

  • To research the major causes of product variability (one cause can impact several measures).

IDENTIFY THE PARAMETERS SETTINGS:

  • Setting process parameters to target the product characteristics.
  • Setting process parameters to compensate for the variation of the raw materials characteristics.

SELECT THE VARIABLES TO BE MEASURED:

  • To monitor product quality (by GPC).
  • To identify anomalies/dysfunctions.

This software allows for selecting variables of a group X to take into account of variations of a group of variables Y.

The main characteristics of GPC PVA software are:

  • •Model of linear relations between X and Y.
  • •Some variables X can be quantitative.
  • •Different criteria of selections depending on use:
    • multiple regression coefficients.
    • residual variance of Y.
    • weighting of deviations in Y.
    • •reduction of the dispersion of Y according to the reduction of the dispersion of X.
  • •Some preferences can be defined on the X plural for integrating the field-knowledge:
    • reliability of measures (presence of information rates).
    • setting variables.
    • low-cost of measure…
  • •Principal variables can be imposed.
  • •The selection can continue in the Y group (relevant interpretations).