Crash narrative classification: Identifying agricultural crashes using machine learning with curated keywords.

OBJECTIVE: Traditionally, structured or coded data fields from a crash report are the basis for identifying crashes involving different types of vehicles, such as farm equipment. However, using only the structured data can lead to misclassification of vehicle or crash type. The objective of the current article is to examine the use of machine learning methods for identifying agricultural crashes based on the crash narrative and to transfer the application of models to different settings (e.g., future years of data, other states). METHODS: Different data representations (e.g., bag-of-words [BoW], bag-of-keywords [BoK]) and document classification algorithms (e.g., support vector machine [SVM], multinomial naïve Bayes classifier [MNB]) were explored using Texas and Louisiana crash narratives across different time periods. RESULTS: The BoK-support vector classifier (SVC), BoK-MNB, and BoW-SVC models trained with Texas data were better predictive models than the baseline rule-based algorithm on the future year test data, with F1 scores of 0.88, 0.89, 0.85 vs. 0.84. The BoK-MNB trained with Louisiana data performed the closest to the baseline rule-based algorithm on the future year test data (F1 scores, 0.91 baseline rule-based algorithm vs. 0.89 BoK-MNB). The BoK-SVC and BoK-MNB models trained with Texas and Louisiana data were better productive models for Texas future year test data with F1 scores 0.89 and 0.90 vs. 0.84. The BoK-MNB model trained with both states' data was a better predictive model for the Louisiana future year test data, F1 score 0.94 vs. 0.91. CONCLUSIONS: The findings of this study support that machine learning methodologies can potentially reduce the amount of human power required to develop key word lists and manually review narratives.

A semi-automated tool for identifying agricultural roadway crashes in crash narratives.

Objective: Crash reports contain precoded structured data fields and a crash narrative that can be a source of rich information not included in the structured data. The narrative can be useful for identifying vulnerable roadway users, such as agricultural workers. However, using the narratives often requires manual reviews that are time consuming and costly. The objective of this research was to develop a simple and relatively inexpensive, semi-automated tool for screening crash narratives and expediting the process of identifying crashes with specific characteristics, such as agricultural crashes. Methods: Crash records for Louisiana from 2010 to 2015 were obtained from the Louisiana Department of Transportation (LaDOTD). Records with narratives were extracted and stratified by vehicle type. The majority of analyses focused on a vehicle type of farm equipment (Type T). Two keyword lists, an inclusion list and an exclusion list, were created based on the published literature, subject-matter experts, and findings from a pilot project. Next, a semi-automated tool was developed in Microsoft Excel to identify agricultural crashes. Lastly, the tool's performance was assessed using a gold standard set of agricultural narratives identified through manual review. Results: The tool reduced the search space (e.g., number of narratives that need manual review) for narratives requiring manual review from 6.7 to 59.4% depending on the research question. Sensitivity was high, with 96.1% of agricultural crash narratives being correctly classified. Of the gold standard agricultural narratives, 58.3% included an equipment keyword and 72.8% included a farm equipment brand. Conclusion: This article provides information on how crash narratives can supplement structured crash data. It also provides an easy-to-implement method to facilitate incorporating narratives into safety research along with keyword lists for identifying agricultural crashes.