Level:
Advanced
TCM Section(s):
10.4. Project Historical Database Management
11.3. Information Management
Venue: 2022 AACE International Conference & Expo
Abstract: The National Nuclear Security Administration (NNSA) collects work breakdown structure (WBS) data for capital asset projects.A WBS allows cost estimators and program managers to track and compare capital acquisition costs across the entire nuclear security enterprise.The difficulty is that implementing the structure across projects grows exponentially in complexity because of varying project scope, contextual changes, and vendor requirements.In this paper, we will demonstrate the capability of natural language processing to identify and classify the WBS elements used for NNSA capital asset projects.
The results indicate that both unsupervised and supervised machine learning (ML) algorithms can obtain a pre-defined classification scheme developed by capital acquisition experts.Multiple ML techniques were employed and are compared based on ease of deployment, computational time to completion, and accuracy with respect to the desired results.These technologies can dramatically alter the cost analyst’s workflow by automating significant and often time-consuming components of his or her work.
TCM Section(s):
10.4. Project Historical Database Management
11.3. Information Management
Venue: 2022 AACE International Conference & Expo
Abstract: The National Nuclear Security Administration (NNSA) collects work breakdown structure (WBS) data for capital asset projects.A WBS allows cost estimators and program managers to track and compare capital acquisition costs across the entire nuclear security enterprise.The difficulty is that implementing the structure across projects grows exponentially in complexity because of varying project scope, contextual changes, and vendor requirements.In this paper, we will demonstrate the capability of natural language processing to identify and classify the WBS elements used for NNSA capital asset projects.
The results indicate that both unsupervised and supervised machine learning (ML) algorithms can obtain a pre-defined classification scheme developed by capital acquisition experts.Multiple ML techniques were employed and are compared based on ease of deployment, computational time to completion, and accuracy with respect to the desired results.These technologies can dramatically alter the cost analyst’s workflow by automating significant and often time-consuming components of his or her work.