Application of machine learning to mapping primary causal factors in self reported safety narratives. (June 2015)

• Record Type: Journal Article
• Authors: Robinson, S.D.; Irwin, W.J.; Kelly, T.K.; Wu, X.O.
• Is Part Of: Safety science. Volume 75(2015)
• Page Start: 118
• Page End: 129
• Publication Date: 2015-06
• Subjects: LSA -- Adaptive taxonomy -- Safety -- Automatic indexing -- Machine learning; Industrial accidents -- Periodicals; Accident Prevention -- Periodicals; Safety -- Periodicals; Travail -- Accidents -- Périodiques; 363.11
• DOI: 10.1016/j.ssci.2015.02.003
• Languages: English
• ISSNs: 0925-7535

Abstract

Highlights: Application of natural language processing techniques to safety report narratives. Compares human-coded primary-cause analysis with narrative context. Develops a statistical metric for analyzing narrative similarity. Automatic categorization supports verification of human coded taxonomies. Filters narrative reports to identify contextual similarity without taxonomy. Abstract: A new method for analysis of text-based reports in accident coding is suggested. This approach utilizes latent semantic analysis to infer higher-order structures between documents and provide an unbiased metric to the narrative analysis process. Results from this study on a small sample of aviation safety narratives demonstrates an unsupervised categorization accuracy of 44% for primary-cause within the existing taxonomy. If provided with a large sample set, the indication is that a significant increase in accuracy is possible along with the possibility of recoding between data sets. Demonstrated is the ability of LSA to capture contextual proximity of a narrative.