ELECTRONIC CATALOG RETRIEVAL SYSTEM


Electronic Catalog

Electronic catalogs contain data of parts and products information used in the heavy electric machinery industry. They contain not only the commercial specifications of parts (manufacturer name , price, etc.), but also the technical specifications (physical size , performance, quality, etc.). The ISO13584 Parts Library, called PLIB for short, is the standard dictionary model and content model for describing both specifications (see Pierra, 1997, 1998).

The dictionary model is a framework for defining the classification and hierarchy of a parts library. A class of parts is characterized by a variety of the attributes of parts, such as manufacturers, product types, performance, etc., and a lower class inherits attributes from its upper classes. On the other hand, the content model is a data scheme for describing both the commercial and technical specifications of parts. The content of each catalog is associated with a class in the dictionary model. Each class and attribute is identified by a unique code called a BSU (Basic Semantic Unit).

Figure 1 shows an example of PLIB catalog dictionaries and contents. A dictionary can be represented in a tree form, as shown in Figure 1(a). Each oval denotes a class, and each edge between ovals stands for an inheritance relation between classes. Each rectangle denotes an attribute, which is associated with a class. For instance, three attributes, LCD, Weight and Battery Life, are associated with the Laptop PC class. The bottom table, shown in Figure 1(b), represents three contents associated with the Laptop class. The contents are defined using five attributes: CPU, Size, LCD, Weight, and Battery Life. The former two attributes are inherited from the PC class, and the latter three attributes come from the Laptop class.

click to expand
Figure 1: Examples of PLIB Catalog Dictionary and Content

An electronic catalog server is an implementation which realizes a marketplace for exchanging catalog data in a PLIB scheme. It is intended to let us retrieve catalog data from electronic servers and then integrate the retrieved data into our own servers, procurement systems, and CAD systems. Currently, major heavy electric machinery companies are building up their own electronic catalog servers with the aim of enclosing their customers, partner groups, and industries.

Integration Problems

This section presents the issues to be resolved for integrating electronic catalog servers.

  1. Retrieval time: There are millions of parts in the heavy electric machinery industry. Catalog service providers have public catalog data for general customers and private catalog data for their specific customers on electronic catalog servers. Thus, it is impossible to manage all the catalog data in a unified way. Although a server should retrieve catalog data in a certain class from other servers if it does not have the data, it generally takes an extremely long time to retrieve the desired catalog data from the mass of catalog data.

  2. System integration: Currently, there is no standard query language and protocol for exchanging PLIB electronic catalog data among electronic catalog servers. The existing servers provide their own distinct ways of querying catalog contents. In order to realize open distribution of catalog contents, each server should be equipped with the query mechanism for any other catalog server if it retrieves and uploads catalog data over the Internet. However, such system integration is very complex and costly.

  3. Various retrieval requests : While the current PLIB standard does not include query ways, various query ways are required for utilizing electronic catalog servers. For example, the way to retrieve catalog data in the specific class in a dictionary model and the way to retrieve any data, including specific manufacturers below a certain class, are required. Furthermore, the PLIB standard allows us to extend a dictionary model to add attributes such as a price for specific customers.




(ed.) Intelligent Agents for Data Mining and Information Retrieval
(ed.) Intelligent Agents for Data Mining and Information Retrieval
ISBN: N/A
EAN: N/A
Year: 2004
Pages: 171

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net