Predictive models for the effect of environmental factors on the abundance of Vibrio parahaemolyticus in oyster farms in Taiwan using extreme gradient boosting

Abstract

This study sought to investigate the effects of environmental parameters on the variation of V. parahaemolyticus in the oyster culture environment in Taiwan. Environmental factors were used to develop predictive models for V. parahaemolyticus concentration in oysters, seawater, and sediment by employing the extreme gradient boosting (XGB) machine learning algorithms. The results showed that XGB capable of predicting the concentration of V. parahaemolyticus in the oysters and seawater, but not for sediment. The relative importance variable analysis showed that V. parahaemolyticus concentration in oysters, seawater, and sediment was dominantly influenced by the variation of sea surface temperature (SST). Increasing wind speed within two days before sampling collection could decrease the number of V. parahaemolyticus in oysters and seawater. The population of V. parahaemolyticus in any type of sample was influenced by the acidity (pH) of seawater. However, the salinity only influenced the concentration of this pathogen in the oysters and sediment, but not in seawater. Thus, monitoring and recording these factors would be useful to predict the level of V. parahaemolyticus in the oyster farms in Taiwan. Findings in this study may be useful in managing the safety of oysters at the farm stage and thus allow the prevention of V. parahaemolyticus infections from eating oysters. • XGB predictive models for V. parahaemolyticus levels in the oyster farms in Taiwan were developed. • V. parahaemolyticus levels were dominantly influenced by sea surface temperature factors. • The effect of a specific factor on the V. parahaemolyticus abundance in samples were evaluated. • Data and models may be used to improve the accuracy of risk assessment model for V. parahaemolyticus .