Interface ThematicClassificationCorrectness

All Superinterfaces:
Element, ThematicAccuracy

@UML(identifier="DQ_ThematicClassificationCorrectness", specification=ISO_19157) public interface ThematicClassificationCorrectness extends ThematicAccuracy
Comparison of the classes assigned to features or their attributes to a universe of discourse. The assignment of an item to a certain class can either be correct or incorrect. Depending on the item that is classified, several data quality measures exist.

Standardized values

In order to achieve well defined and comparable quality information, it is recommended to report data quality using quality measures listed in ISO 19157 annex. The following table provides a summary; see ISO 19157 for more complete descriptions and examples. All identifiers should be in "ISO 19157" namespace.
Standardized values derived from ISO 19157
Identifier Name of measure Basic measure Parameters Value type
60 number of incorrectly classified features error count Integer
61 misclassification rate error rate Real
62 misclassification matrix Matrix
63 relative misclassification matrix Matrix
64 kappa coefficient n (number of classes) Matrix

Note: ISO 19157 declares the misclassification matrix value type as "Integer" or "Real" associated with ValueStructure.MATRIX. For an object oriented language like Java, a more natural approach is to use an object of specific type for the value.

Definitions:

  1. Number of incorrectly classified features
  2. Number of incorrectly classified features relative to the number of features that should be there.
  3. Matrix that indicates the number of items of class (i) classified as class (j).
  4. Matrix that indicates the number of items of class (i) classified as class (j) divided by the number of items of class (i).
  5. coefficient to quantify the proportion of agreement of assignments to classes by removing misclassifications.
Since:
2.0