Classification

With the menu items of the classification, it can be displayed in detail and also changed. The display depends on the feature type (variable / attributive). Please note that changing the classification in the standard is a temporary setting and can have an effect on the CHI² test when using test procedures for non-normal distributions. Therefore, it is not recommended to change the classification permanently.

General configuration
The "Graphical settings" tab provides all the options for modifying the graphics, in colour, font or content.
Link to: Q-DAS Graphics - General Configuration

Working with the graphics
The "Part / characteristic" tab provides options and functions for working with graphics. These include various data selection options for display and evaluation, as well as various configuration options for the displaying multiple characteristics.
Link to: Working with the Q-DAS Graphics

This topic describes the special settings that only apply to this graphic.

1 Classification
- 1.1 Classification model
- 1.2 Classification
- 1.3 Pie chart
Topic in PFD format

1 Classification

Under the menu item of classification, the classification defined by the strategy can be viewed or also modified (in the case of variable characteristics).

When changing the classification, note that this is a volatile setting by default.

In almost all newer strategies, there are other distribution forms besides the default of normal distribution, which are given and tested as the first idea. In this case, the "CHI² test for any distribution" must be activated. Together with this, the classification must be set to "Classification under consideration of natural resolution". For this reason, it is no longer recommended to set or save a different classification, as this could have negative effects on the CHI² test.

1.1 Classification model

This setting applies temporarily to the respective characteristic of the currently loaded file.

John/Q-DAS

The number of classes is between the square and cube root of n (n = subgroup size). Class boundaries are formed as smoothly as possible.

DIN 55302-1/Q-THIS

Classification model in which the requirement for the minimum number of classes is only met from n = 100. For smaller subgroup sizes, the number of classes results from the square root of n.

DIN 55302-1

The minimum number of classes is fixed at 10 even with a subgroup size of n < 100.

Classification with consideration of resolution

In this classification model, the class width is an integer multiple of the resolution.

This is important when using the CHI² test.

Sturges/CNOMO

Model based on the French CNOMO standard.

Fixed number of classesoutside tolerance

With this model you can define how many fixed classes lie outside the specification limits. After selecting this option, the two additional areas "Class Limits" and "Alternative Classification Model" appear in the dialogue.

Under "Class limits" you can specify the fixed number of classes within/outside the specification limits.

In addition, you still have to select an alternative classification model if the fixed number of classes cannot be used.

Fixed number of classes within tolerance

With this model you can define how many fixed classes lie within the specification limits. After selecting this option, the two additional areas "Class Limits" and "Alternative Classification Model" appear in the dialogue.

Under "Class Limits" you can preset the fixed number of classes within the specification limits.

In addition, you still have to select an alternative classification model if the fixed number of classes cannot be used.

Enter class boundaries

By activating this option field, you can specify a class limit individually for the respective characteristic. The two areas "Class limits" and "Rounding" appear.

Under "Class limits" you can enter any lower/upper class limit.

The two buttons (min/max) are used to enter the highest or lowest values of the respective characteristic into the mask.

Use the two buttons LSL and USL to enter the respective specification limits of the characteristic from the characteristics mask.

In addition, you can specify the number of classes (this must always be done manually).

Under "Rounding" you can specify to which class the respective measured value is to be assigned based on the natural resolution (decimal places).

1.2 Classification

The individuals from the histogram that have been combined into classes can be displayed with this option. The classes with a consecutive number are listed here with their class boundaries and the associated absolute and relative frequencies. For visualisation, the frequency is additionally displayed graphically with a bar chart.

The graphic is pre-configured for the different feature types.

Variable characteristics

x'_j-1 = lower class limits

x _j = class centre

x' _j = Upper class limit

w _j = class width

n _j = Absolute frequency

h _j = Relative frequency

G _j = Absolute cumulative frequency

F _j = Relative cumulative frequency

The class display thus corresponds to a rotated histogram:

Attributive characteristics

For attributive characteristics, the two classes i.O. and n.i.O. are represented in the classification.

Ordinal and nominal characteristics

For nominal characteristics, the class representation is limited to the class description, the absolute and relative frequency and the graphical representation. The order of the classes follows the rank entered in the catalogues.

1.3 Pie chart

This option allows you to visualise the distribution of the individual measured values within the different classes using a pie chart. Even though this is also possible for variable characteristics, the pie chart was initially created only for attributive, ordinal and nominal characteristics.