Knowledge Discovery in Transportation Databases (KDITD)

Over the years, Department of Transportation (DOT) agencies along with university laboratories have produced a large number of transportation system databases that cover a wide spectrum ranging from crash-traffic historical databases to pavement-material performance databases. The main goal of the knowledge discovery process, via the use of statistically-based, artificial intelligence-based or hybrid techniques, is to extract the knowledge buried within these databases. Knowledge discovery via statistic ally-based techniques is widely used. However, these techniques have severe limitations and constrains in extracting the knowledge due to the complexity of these databases. On the other hand, artificial intelligence-based methods such as artificial neural networks, fuzzy logic, and other various forms of data mining procedures offer a more efficient methodology for knowledge discovery. For example, artificial intelligence-based (AIB) methods are not constrained by the required functional form that is typically needed to be defined in advance for statistically-based (SB) methods. Appropriately combining the best features of SB and AIB methodologies can a yield a far more superior hybrid knowledge discovery (HKD) approach. Utilizing such hybrid approach can efficiently extract t he important features (i.e., useful knowledge) hidden in the complex transportation system databases. In this research study, we are proposing to appropriately combine the diverse expertise of the research team in order to develop an efficient HKD approach. The developed HKD approach will then be used to extract the hidden features within a wide spectrum of databases ranging from crash-traffic-driver related databases to pavement-material-performance related databases. It is expected that this research will help advance the proper utilization of the newly created information-based science (IBS) in transportation engineering.

• Record URL:
  http://transport.ksu.edu/files/transport/imported/newsletters/KSUTC-1.pdf
• 

Language

• English

Project

• Status: Active
• Funding: $180000.00
• Contract Numbers:

  RE-040-05; HPD-R043

  KSUTC-08-8

• Sponsor Organizations:

  Kansas State University Transportation Center

  Kansas State University
  Department of Civil Engineering
  Manhattan, KS  United States  66506
• Project Managers:

  Stokes, Robert

• Performing Organizations:

  Kansas State University Transportation Center

  Kansas State University
  Department of Civil Engineering
  Manhattan, KS  United States  66506
• Principal Investigators:

  Peric, Dunja

  Dissanayake, Sunanda

  Najjar, Yacoub

• Start Date: 20050810
• Expected Completion Date: 0
• Actual Completion Date: 20120531
• Source Data: RiP Project 15286

Subject/Index Terms

• TRT Terms: Artificial intelligence; Databases; Expert systems; Fuzzy logic; Knowledge; Knowledge acquisition (Expert systems); Machine learning; Neural networks; Research projects
• Subject Areas: Data and Information Technology; Highways;

Filing Info

• Accession Number: 01462763
• Record Type: Research project
• Source Agency: Kansas State University Transportation Center
• Contract Numbers: RE-040-05; HPD-R043, KSUTC-08-8
• Files: UTC, RIP
• Created Date: Jan 3 2013 2:09PM