Calculation Information

The cost calculation is based on the correlation and regression analysis. These are mathematical-statistical methods used for identifying and describing the dependencies between two or more attributes. While the regression analysis explores the nature of the relationship between the attributes investigated, the correlation analysis to draw conclusions on the strength of the dependency.

The calculation is based on the assumption of an exponential relationship between input parameter and costs, i.e. a linear correlation. It means that the results of a subsequent draft are calculated by the size of the basic draft

with

a cost function.

The costs are computed as follows with dependence on the selection of the search criteria and the result list:

• If no influence variables exist or none are selected, the costs are given as the arithmetic mean value of all the found and marked components.

• If influence variables exist and have been selected:

  • Functional dependencies between the selected influence variables and the costs of all found and marked components are examined as well as the part spectrum, although only by extrapolation. (Influence of the Forecast - Part lies beyond the area of smallest and biggest influences from the reference part quantity)
  • If no dependencies are detectable, the costs are also determined as arithmetic mean value.
  • If dependencies are found, the corresponding influence variables will be activated. If no extrapolation is shown, the costs will be given as a result of the regression analysis and the influence variables applied to the regression analysis will be marked.

The interface displayed will vary depending on the status of the calculation.

• "Correl.": Describes the empirical correlation coefficient. This is a means of measuring the strength and direction of the functional dependency. Resolution has a range between -1 and 1:
  • Values > 0 imply large influence variables predominantly correspond to higher costs.
  • Values < 0 imply large influence variables predominantly correspond to lower costs.
  • Values approaching 1 or -1 imply a strong functional dependency.
  • Values approaching 0 imply a weak functional dependency.

• "Stat.": Refers to the statistical spread of the empirical regression function. It is a means of measuring the extent to which the regression function fits the scatter plot of the characteristic.

  • Values approaching 0 indicate a good adjustment.
  • Values approaching infinity indicate a bad adjustment.

• "Res.": Refers to the resolution of the regression analysis. This is a measure of how well the regression function describes the relationship between the influence variable and the costs. Resolution has a range between 0 and 1.

  • Resolutions approaching 0 indicate a bad approximation.
  • Resolutions approaching 1 indicate a good approximation.

• "Exp.": Describes the cost exponent indicating the potency in which the costs generally grow to the influencing variable scale .