TIES Living Lab
Artificial Intelligence Project Data Mining
In most infrastructure and construction projects, the actual cost is often significantly higher than estimated one. Uncertainty surrounding cost planning and especially in the initial stages of projects, remains as one of the thorniest issues in public projects, affecting the quest to improve efficiency and productivity. Enabling research and understanding around cost planning and estimating of construction investment is a key capability in driving these efficiency savings, informing early business case decisions, and demonstrating value for money through benchmarking.
However, the reputation of the transport infrastructure sector’s cost planning and estimating capability is not strong, its connection with research disjointed, at best and its benchmarking capability is incoherent. Shortcomings in this capability also have a direct impact on the reputation of clients, stakeholders and the industry as a whole.
TIES Project 2 Artificial Intelligence and Data Mining seeks to develop an Artificial Intelligence (AI) driven system that facilitates cost (and quantities where available) data mining for benchmarking and analysis. This project developed AI methods and techniques, to enable the efficient translation of cost information produced in different formats and using different standards and structures into a consistent structure according to the International Cost Measurement Standard (ICMS) to support benchmarking. This enables comparable cost metrics and benchmarks to be consistently determined and also provides a robust basis for other metrics based on costs and quantities, such as environmental benchmarks and whole life analysis. By facilitating the semi-automated classification of cost data into ICMS, this work also helps to support the estimating process, according to ICMS structure, and enhance the harmonisation of cost reporting within and across organisations, and the supply chain. It delivers consistency in cost reporting and monitoring of projects, as well a aid to achieve a better value for money, and best practices in benchmarking.
- AI approaches and techniques to classify structured and unstructured cost data descriptions and variables into a common ICMS format.
- A semi-automated system for extraction, identification, translation and mapping of cost related data into a standardised ICMS schema for further benchmarking analysis.
- Project risk classification based on natural language processing.
- Final project cost prediction using historical preliminary project data.