Oldroyd, R.A.; Morris, M.A.; Birkin, M. Predicting Food Safety Compliance for Informed Food Outlet Inspections: A Machine Learning Approach. Int. J. Environ. Res. Public Health 2021, 18, 12635. https://doi.org/10.3390/ijerph182312635
Oldroyd, R.A.; Morris, M.A.; Birkin, M. Predicting Food Safety Compliance for Informed Food Outlet Inspections: A Machine Learning Approach. Int. J. Environ. Res. Public Health 2021, 18, 12635. https://doi.org/10.3390/ijerph182312635
Oldroyd, R.A.; Morris, M.A.; Birkin, M. Predicting Food Safety Compliance for Informed Food Outlet Inspections: A Machine Learning Approach. Int. J. Environ. Res. Public Health 2021, 18, 12635. https://doi.org/10.3390/ijerph182312635
Oldroyd, R.A.; Morris, M.A.; Birkin, M. Predicting Food Safety Compliance for Informed Food Outlet Inspections: A Machine Learning Approach. Int. J. Environ. Res. Public Health 2021, 18, 12635. https://doi.org/10.3390/ijerph182312635
Abstract
Consumer food environments have transformed dramatically in the last decade. The number of food outlets has increased, and a large proportion of the UK population now purchase food from a takeaway or restaurant at least once a week. Despite these developments, national spending on food control has reduced and many Local Authorities struggle to meet health inspection targets. This work presents a data driven approach to enhance current inspection processes with a view to reduce consumer risk of foodborne illness whilst eating outside the home. We explore the utility of three machine learning algorithms to predict non-compliant food outlets in England and Wales as defined by Food Hygiene Rating Scheme scores >= 2. Using socio-demographic, business type and urbanness features we experiment with under and over sampling strategies at five ratios to address problems of class imbalance in the dataset prior to analysis. We find that Synthetic Minority Over Sampling Technique alongside a Random Forest algorithm with a 1:1 sampling strategy provides the best predictive power. Our final model retrieves 84% of total non-compliant outlets in a test set of 92,595 (sensitivity=0.843, specificity=0.745, precision=0.274). We discuss the utility of machine learning algorithms to prioritise high risk establishments for inspection by Local Authority officials and make recommendations for weighting outcomes to improve their appropriateness in an applied setting.
Biology and Life Sciences, Food Science and Technology
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
