Database mining in the Northern Ireland Housing Executive

Abstract

Describes some work being carried out by the Database Mining Research Group (DMRG) within the School of Information and Software Engineering, at the University of Ulster, Jordanstown. In particular the authors outline how they have been investigating the use of data mining techniques on data sets from the Northern Ireland Housing Executive (NIHE). NIHE are responsible for the management of the public sector housing stock. They are the largest such body within the European Union, and are acknowledged as leaders within their field in the use of information technology. Within NIHE there are numerous databases in use. These fall into two main categories. There are the day to day, operational databases such as the Prawl rent accounting system and the Repairs maintenance management system. Such databases are used both for transaction processing and management information purposes. The other main category of database in use are the various data sets used by the Research Department. These are used for analysis that feeds into the formulation of policy. With such a large and growing amount of data within their organisation, NIHE are interested in techniques that will help them make better use of it. It was decided in the first instance to use the House Condition Survey data set as the target data set. Several data mining techniques being explored by the DMRG are being evaluated using this data. The techniques in question are algorithms for the discovery of classification and association rules. The authors are also interested in sequential or temporal rules but the work on this is less advanced.