Automatically classify Bills of Quantities descriptions into the International Construction Measurement Standard (ICMS) using state-of-the-art machine learning.
Test the Classifier
Enter a construction item description to see its ICMS classification.
Classifying...
Classification Result
ICMS Code:
Level 2:
Level 3:
Level 4:
Error
Batch Classification
Upload a CSV file to classify multiple items at once. The first column should contain item descriptions. Maximum 100 rows.
The International Construction Measurement Standard (ICMS) is a globally consistent framework for measuring and reporting construction project costs. Developed by leading construction industry organisations, ICMS provides a standardised approach to cost classification, enabling better cost comparison and benchmarking across projects and countries.
Why ICMS Matters
Consistency: Enables accurate cost comparisons between projects
Efficiency: Reduces time spent on cost classification and analysis
Global Standard: Recognised and used internationally
The ICMS Coalition
The ICMS Coalition comprises 49 globally prominent organisations working collaboratively to develop and maintain the ICMS standards. With the built environment responsible for approximately 40% of global carbon emissions, the Coalition has developed ICMS 3—a world first that provides a globally consistent method for both cost and carbon life cycle reporting. This advancement empowers professionals to deliver construction projects whilst managing both financial costs and carbon emissions effectively, supporting the construction sector's decarbonisation efforts.
ICMS Structure
ICMS organises construction costs into a hierarchical structure with four levels:
Level 1: Project or subproject
Level 2: Major cost categories (e.g., Capital Construction Costs)
Level 3: Cost groups (e.g., Substructure, Superstructure)
Level 4: Detailed cost categories (e.g., specific elements and components)
About This Tool
This classifier uses machine learning trained on construction item descriptions from UK infrastructure projects to automatically assign ICMS codes to Bills of Quantities descriptions.
The model employs a Random Forest algorithm with TF-IDF vectorisation, optimised for the short, technical language typical of construction documentation. Research has shown that for BoQ text, information is embedded in local key features rather than long-range dependencies, making this approach suitable for the task.
Research & Development
This tool was developed as part of the TIES Living Lab programme, specifically within Work Package 2: AI for Data Mining. The research and implementation were conducted by the Centre for Advanced Built Environment Research (CABER) at UWE Bristol.
Published Research:
Deza, J.I., Ihshaish, H. and Mahdjoubi, L. (2022).
"A Machine Learning Approach to Classifying Construction Cost Documents into the International Construction Measurement Standard"
TIES Living Lab is a transformative collaboration of 25 partners working with Government, i3P and the Construction Innovation Hub to deliver innovation in transport infrastructure through data, technology and Modern Methods of Construction.
