KNOWLEDGE DISCOVERY IN DATABASES AND PAVEMENT MANAGEMENT SYSTEM INTEGRATION FRAMEWORK

*Transportation Information Technology Expert

**Cairo University – Egypt

***Fayoum University – Egypt

Databases are considered a key element in any management system. In Pavement

Management Systems (PMS), database interpretation is a crucial part throughout the

management process. The amount of data collected throughout developing the PMS database

is recognized to be huge enough needing computer assistance for proper interpretation.

Consequently, many studies have been performed to include the automated bay to PMS

starting from primitive computer programs and ending with Expert Systems. New techniques

and tools have been developed to provide a pave for the gap that is expanding between the

analysis method used and the amount of data available. Through the past ten years,

Knowledge Discovery in Databases (KDD) has gained a special interest for its powerful way

of handling huge amount of data not only in analysis but also for knowledge extraction. The

objective of this paper is to implement the KDD techniques on the PMS database through a

case study for predicting the Remaining Life (RL) of a pavement. As a mater of fact,

discovering new relationships between RL and the different pavement characteristics is of

major interest. To achieve this objective, a framework has been constructed, which consists

of four modules: case articulation, data preparation, data mining, and knowledge

interpretation. Data mining was applied using the GeneMiner-2001 software. The knowledge

or the extracted pattern indicated that there is a strong correlation between the RL of a

pavement and the road width. On the other hand, a Decision Support System (DSS) was

developed by visual basic programming language to predict the RL of a pavement section.

The prediction and validation process was performed. The results from the DSS program is in

a Fuzzy style (ranges). This is due to the algorithm of the software (GeneMiner) used in

mining process. This attempt is considered influential for enhancement of PMS.