Predicting construction safety outcomes with machine learning
This is the free video Predicting construction safety outcomes with machine learning that can be downloaded, played and edit with our RedcoolMedia movie maker MovieStudio free video editor online and AudioStudio free audio editor online
VIDEO DESCRIPTION:
Play, download and edit the free video Predicting construction safety outcomes with machine learning.
NLP is used to extract attributes from a large dataset of injury reports and machine learning models are trained to predict independent safety outcomes (human annotations).
Four safety outcomes are predicted: injury severity, injury type, bodypart impacted, and incident type.
Random Forest, XGBoost and linear SVM are compared and combined through model stacking.
Results show that attributes are highly predictive, even for the injury severity outcome.
Read the paper (accepted for publication in Automation in Construction):
- AI-based Prediction of Independent Construction Safety Outcomes from Universal Attributes arxiv.org/pdf/1908.05972.pdf
Visit safetyfunction.com for more.
Safety AI: state-of-the-art artificial intelligence for your construction safety data.
Download, play and edit free videos and free audios from Predicting construction safety outcomes with machine learning using RedcoolMedia.net web apps