Knowledge discovery and pavement performance: intelligent data mining

The main goal of the study was to discover knowledge from data about asphalt road pavement problems to achieve a better understanding of the behavior of them and via this understanding improve pavement quality and enhance its lifespan. Four pavement problems were chosen to be investigated; raveling of porous asphalt concrete (PAC), cracking of dense asphalt concrete (DAC), rutting of dense asphalt concrete, and determination of the stiffness of cement treated bases (CTBs). Determination of the stiffness of the cement treated base layer stiffness is not an easy task. Therefore, a tool which can accurately calculate the stiffness of such base layers is desirable. The data for the stiffness of CTBs was simulated using the multilayer linear-elastic computer program BISAR. During preparation of the data, the determination of outliers was a challenging task. Due to the low number of data points available for raveling, cracking, and rutting (in one case around 70 data points), an extensive variable selection was performed using eight different methods: decision trees, genetic polynomial, artificial neural network, rough set theory, correlation based variable selection with bidirectional and genetic search, wrappers of neural network with genetic search, and relief ranking filter. For development of models (data mining) from the mentioned data, four machine learning based techniques were employed. This study resulted in 20 intelligent models for the mentioned four problems.

Language

  • English

Media Info

  • Pagination: 1 file

Subject/Index Terms

Filing Info

  • Accession Number: 01384167
  • Record Type: Publication
  • Source Agency: ARRB
  • ISBN: 9789085702788
  • Files: ATRI
  • Created Date: Aug 22 2012 4:15PM