Analyzing Asset Management Data Using Data and Text Mining

Details

• Creators:
  Williams, Trefor ; Halling, Marv
• Corporate Creators:
  Rutgers University. Department of Civil & Environmental Engineering
• Corporate Contributors:
  United States. Dept. of Transportation. Research and Innovative Technology Administration
• Subject/TRT Terms:
  • Asset management
  • Data analysis
  • Data mining
  • Mathematical prediction
  • Algorithms
  • Construction projects
  • Costs
  • Research Hub
  • Text mining
  • Construction
  More +
• Publication/ Report Number:
  CAIT-UTC-031
• Resource Type:
  Tech Report
• Geographical Coverage:
  California ; United States
• TRIS Online Accession Number:
  01570305
• Edition:
  Final Report
• Corporate Publisher:
  Rutgers University. Center for Advanced Infrastructure and Transportation
• Abstract:
  Predictive models using text from a sample competitively bid California highway projects have been used to predict a construction projects likely level of cost overrun. A text description of the project and the text of the five largest project line items were used as input. The text data were converted to numerical attributes using text-mining algorithms and singular value decomposition. Two models were produced. The first used only the text description as input, while the second combined the text data with the numeric value of the low bid. Classification models were produced using the K-Star classification algorithm. Modeling results indicated information in the textual descriptions is related to the projects level of cost overrun.
• Format:
  PDF
• Funding:
  DTRT12-G-UTC16
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/28672/dot_28672_DS1.pdf
• File Type:
  [PDF - 475.62 KB]
• Collection(s):
  University Transportation Centers ; US Transportation Collection
• Main Document Checksum:
  urn:sha256:41c8096d32e7ec4a6f33c614d87da1d0a3fe7f0feb940f5cb986ca3b4fca6eaf