Article Preview
TopIntroduction
The traditional Data Warehouse supports the cleaning, transformation and loading of corporate data into libraries. The corporate data contained in warehouse may be of multiple granule size to make it capable of handling queries of versatile nature.
According to Husemann, Lechtenbörger, and Vossen (2000), a data warehouse design is commonly supported by a conceptual data model called multidimensional model by which users could view data from different dimensions necessary for analysis purposes. In multidimensional model, data are represented in terms of facts and dimensions where each fact is associated to multiple dimensions. In this manner, facts are the focus of interest by which they are analyzed through the quantifying context stored in measures and the qualifying context determined through dimension levels. Categorizing data along dimensions is a mean to organize them into hierarchical levels so that data can be viewed from their finer to coarser granularities as per Agrawal, El Abbadi, Singh, and Yurek (1997). Data warehouses store data and can be a source of knowledge but do not store knowledge directly. Knowledge is presented in the DW in the form of analysis reports, stored statically, the value and importance of which may vary from time to time. To make this knowledge available, Metadata is defined which describes data attributes, transformations and aggregation levels. Metadata helps socialize the Data Warehouse to the knowledge workers so that they can discover information contained and often hidden within the data.
Applications of artificial intelligence (AI) technology in the form of knowledge-based systems within the context of database design have been extensively researched particularly to provide support within the conceptual design phase can be found in Phipps and Davis (2002), Hahn, Sapia, and Blaschka (2000), and Sitompul and Noah (2003, 2005). However, a similar approach to the task of data warehouse design has yet to be seriously initiated. A design methodology is proposed by Sitompul and Noah (2006) for conceptual data warehouse design called the transformation-oriented methodology, which transforms an Entity-Relationship (ER) model into a multidimensional model based on a series of transformation and analysis rules.
The next generation Data Warehouse systems are constructed as Intelligent Data Warehouses (IDW). Bramblett (2002) says that IDW’s are Data Warehouses that are managed as active data sources by a Knowledge Warehouse (KW). The Knowledge Warehouse applies knowledge objects that are created and controlled by software engines using expert system models. Knowledge objects supply the rules, methods and procedures in a reusable way and are activated to not only link Data Warehouse data with a specific business process, but to re-package data so that the IDW can manage the business. The existing enterprise-wide information delivery systems provided in a data warehouse can be leveraged and extended to create a knowledge warehouse detail can be found in Nemati, Steiger, Iyer, and Herschel (2002). This warehouse can be used as a clearinghouse of knowledge to be used throughout the organization by the knowledge workers to support their knowledge intensive decision-making activities. The KW can also evolve over time by enhancing the knowledge it contains.
The Data Warehousing System, which began as a low volatility system, is now a system that may integrate DSS, batch and OLTP processing, and that therefore may incorporate considerable volatility. According to Firestone (2000), the new generation Data Warehousing raises the following issues.
- •
Achieve dynamic integration.
- •
Comprehensively integrate and support knowledge production.
- •
Store knowledge for high capability decision support.
- •
Efficiently deliver tactical decision support using volatile data stores.
- •
Integrate ERP systems.
- •
Integrate increasingly varied business process engines.