This paper describes the investigation of data mining Association Rules within a highway database. Association Rules based data mining is widely used within the insurance, banking and retail industries. It is equally applicable to the analysis of the large volume of data collected as part of highway mechanical data collection. Association rules are to be found in data when items exist together in recurring groups. The database reported on in this paper is one that contains mechanically collected High Speed data (roughness, rutting and texture), which is used in the management of a portion of New Zealand's highway network. This paper reports the outcome of the project to develop an understanding of the relevance of mechanically collected data as an indicator of road condition or maintenance "need". To date this understanding has not been well developed amongst highway maintenance managers. (Author/publisher) For the covering entry of this conference, please see ITRD abstract no. E208431.
Abstract