Methods available for analysing data include regression and trend compensation.
Table of Contents
1 Regression functions in the value chart
In the group "Regression" in the tab Part/Characteristic you get various functions for regression and trend compensation.
1.1 Regression
After selecting this option, marker bars are displayed in the graph. With these bars you limit the range of values for which a regression line is to be created. You move the bars with the mouse. Two yellowish fields are now displayed below the graphic, indicating the exact position of the bars. By selecting these fields, a fine adjustment is possible by input or via the arrow keys. In addition, you can see the equation of the regression line and the regression coefficient in the blue fields above the value curve.
1.2 Trend compensation
With the trend compensation function, it is possible to look at how "good" a process would be if it were not for the technically induced trend. Using the example of a trend process, where the intervention always starts more in the area of the upper tolerance:
Now it is to be considered whether the trend alone is only the "problem of the process". For this purpose, the trend compensation is called up and confirmed directly. A manual range definition does not take place in this example. The ranges found by the basic setting are thus accepted.
The following picture then emerges: a separate regression line is shown for each of the automatically found areas.
Directly calculated is the "Trend Compensated Capability Index", which can be found directly on the Form 5.
This "what-if index" is never calculated automatically, and is never part of the "requirement". It is only intended to help with analysis.
A rough pictorial explanation of the calculation of this index: Each individual range with its regression line is rotated and shifted so that the regression lines lie horizontally on the centre of tolerance. The calculation of the trend-compensated index is done with these "shifted" values.
| Regression This document describes how to use the regression and trend compensation functions. |
English:PDF | German:PDF |
