Automatic Identification Of Incidents Involving Potential Serious Injuries And Fatalities (psif)

This article discusses the use of transformer-based deep learning models for risk identification in safety incident management. The authors utilized traditional machine learning methods to create a vector representation of incident records from over 600 customers, resulting in a dataset of 2.2 million records. The data includes diverse columns such as textual, quantity-based, categorical, and date and time columns, with a small portion manually flagged as PSIF records.