Enriching hierarchies in multidimentional model of data warehouse using wordnet

Abstract

In Multidimensional Model of Data Warehouse (DW), dimension tables hierarchies play important roles in Online Analytical Processing (OLAP) analysis. Hierarchies in dimension table allow one to navigate from a detailed to a more general level of data. Data that are analyzed by using many levels of aggregation produce more meaningful analysis. Therefore, it is better for OLAP if dimension tables have rich hierarchies. However, a requirement of data to obtain rich hierarchies might not be captured by existing Online Transactional Processing (OLTP) systems. It is because OLTP usually have limited number of hierarchies’ levels in their database schema. As a result, DW schema design could not have hierarchies’ levels beyond what OLTP schema has. This study proposed a method to capture required hierarchies in operational data by using a lexical ontology’s semantic relations among concepts. By identifying hidden hierarchies in operational data, new levels of aggregation can then be added into dimension table. As a result, dimension table hierarchies become richer. A prototype is developed to demonstrate the concept employed in this research work. The purpose of this study is to show that, process of enriching dimension hierarchies can be automated by extracting knowledge stored in a lexical ontology such as WordNet.