Facet classification

from Wikipedia, the free encyclopedia

A facet classification (also analytical-synthetic classification ) is a classification system in which the objects of a knowledge area are not integrated into a relatively inflexible tree structure, as is the case with purely hierarchical systems. Instead, an object is classified by assigning several independent terms. To do this, several views ( aspects or so-called facets ) are introduced, which represent a certain subdivision aspect (example: "gender"). The specific characteristics of a facet are called foci or simple classes (example: “male”, “female”). Foci of several facets are typically assigned to an object, which is why it is a polyhierarchical classification system. However, a focus only ever belongs to exactly one facet. The foci of a facet can have a hierarchical structure. Several foci of a facet can also be assigned to an object; a facet classification can thus be used in a polydimensional manner.

The classification of an object results from the combination of the assigned foci, which is referred to as post-combination - in contrast to a pre-combination , in which the combination takes place when the terms of the classification system are formed. For the notation of the object classification, the identifiers of the foci are connected, which results in the overall notation .

The best known facet classification is the colon classification . For the area of architecture and construction there is the BRD / SfB system , which is structured in the form of a facet classification.

Advantages and disadvantages

A great advantage of a facet classification compared to pre-combined classification systems is the greater flexibility and expandability. In particular, the number of classes required for a facet classification is considerably lower. For example, the number of required classes could be reduced by a factor of around 40 if the “literature” area of ​​the pre-combined universal decimal classification were mapped in the sense of a facet classification . This means that the effort to maintain the classification catalog is significantly lower.

In contrast to facet classification, precombined and especially mono-hierarchical classification systems specify the access path according to the top-down principle. They are therefore particularly suitable for cases in which there is a typical access path. A predefined access path is also an orientation aid for researchers. In contrast, facet classifications are particularly suitable when access is typically to take place from different perspectives.

In the case of electronic access, the design of an intuitively operable user interface when using a facet classification is usually more demanding compared to a pre-combined classification. Systems that support facet classification typically also require more resources when carrying out a search, because the various search terms are only combined when the search is carried out.

example

Possible classification systems for the subject area "shoe trade"

In the example on the right, two variants of classification systems are outlined, which could represent an excerpt from the subject area “shoe trade”. The upper one represents a mono-hierarchical catalog in which a pre-combination is made. The lower variant is a facet classification. To the right of this is shown how the same specific object - a Budapest , a men's shoe model, size 49 - would be classified in both systems. In the pre-combined system, this object would only be assigned to the “oversized men's street shoe” class. In a facet classification, on the other hand, there would be three, namely “men's shoe”, “street shoe” and “oversize”.

If you are interested in a specific shoe model, the pre-combined system enables you to find the most suitable, detailed knot from top to bottom along the tree. If, on the other hand, you are interested in how large the total range of oversized shoes is, you would have to consider four different classes in the pre-combined system. The difference between the two systems becomes particularly clear if one imagines that a differentiation according to upper material should be added to the classification system. With the pre-combined system, this differentiation would have to take place at many different points in the tree. With the facet classification, on the other hand, only a new facet with the foci “leather”, “textiles”, “synthetic” etc. would have to be introduced. If further differentiations are introduced, for example if orthopedic shoes are to be added to the range, the number of classes required in the pre-combined system grows exponentially .

literature

  • Jutta Bertram: Introduction to the content indexing. Ergon Verlag, Würzburg 2005, ISBN 3-89913-442-7 .
  • Wolfgang G. Stock , Mechthild Stock: Knowledge representation. Oldenbourg Verlag, Munich 2008, ISBN 978-3-486-58439-4 .
  • David Batty: Origins and development of faceted classification . CDB Enterprises, 1983.

Web links

Individual evidence

  1. ↑ Organize, search, find, building information with the BRD / SfB system, 1st edition © 1978 Verlagsgesellschaft Rudolf Müller GmbH, Cologne-Braunsfeld, ISBN 3-481-17951-0
  2. ^ Wolfgang G. Stock, Mechthild Stock: Knowledge Representation. Pp. 271-276, see literature
  3. ↑ a b Jutta Betram: Introduction to the indexing of the content. Pp. 173-181, see literature
  4. ↑ Kerstin Reinhard: Comparative usability evaluation to determine best practice solutions for facetted navigation. Hildesheim University Foundation, Institute for Information Science and Language Technology, Hildesheim 2010, p. 30f ( online )
  5. ↑ Jutta Betram: Introduction to the indexing of the content. Pp. 168-173, see literature