Predicting Seriousness of Injury in a Traffic Accident: A New Imbalanced Dataset and Benchmark

TL;DR

This paper presents a new imbalanced traffic accident injury dataset, along with baseline machine learning models, to improve injury severity prediction and stimulate research on imbalanced data handling.

Contribution

It introduces a novel, highly imbalanced traffic injury dataset and baseline models, addressing data incompleteness and providing benchmarks for future research.

Findings

The dataset reveals significant data imbalance and missing attributes.

Baseline neural network and reinforcement learning models establish reference points.

The dataset and models can facilitate research on imbalanced injury prediction.

Abstract

The paper introduces a new dataset to assess the performance of machine learning algorithms in the prediction of the seriousness of injury in a traffic accident. The dataset is created by aggregating publicly available datasets from the UK Department for Transport, which are drastically imbalanced with missing attributes sometimes approaching 50\% of the overall data dimensionality. The paper presents the data analysis pipeline starting from the publicly available data of road traffic accidents and ending with predictors of possible injuries and their degree of severity. It addresses the huge incompleteness of public data with a MissForest model. The paper also introduces two baseline approaches to create injury predictors: a supervised artificial neural network and a reinforcement learning model. The dataset can potentially stimulate diverse aspects of machine learning research on imbalanced datasets and the two approaches can be used as baseline references when researchers test more advanced learning algorithms in this area.