Answers are grounded in this paper and cite their sources.
Asian Citrus Psyllid, Diaphorina citri Kuwayama (Liviidae: Hemiptera) is a threatening and serious pest of Rutaceae family plants especially citrus and it vectors a phloem vessel-dwelling bacterium Candidatus Liberibacter asiaticus which is causative pathogen of a serious disease called citrus greening disease or Huanglongbing (HLB). This citrus disease is responsible for significant economic losses around the globe and is a major bottleneck in the export of citrus fruits globally. Beneficial insects as Biological Control Agents (BCAs) can suppress pest outbreaks by naturally reducing pest density. Tracking population trends of pest-natural enemies can assist farmers in taking informed decisions regarding pesticides applications. In this research, population trends of ACP-Natural enemies on citrus cultivars by using various Machine learning techniques were monitored. Researchers correlated observed values of natural enemies of ACP with different weather variables and host plant phenological patterns and found their cumulative effect on predators’ population fluctuating trends. Several regression-based approaches were compared based on their predictive ability and found that blended ensemble learning approach combining the characteristics of Random Forest Regressor and Extreme gradient Boosting regressor is an effective approach to model population dynamics of natural enemies of ACP. The regression-based approaches used in this study were Random Forest Regressor, Multiple Linear Regression, Deep Neural Network and an ensemble model because all these techniques have potential to model time-series prediction problems. After model training on training splits of four datasets, predictive performance of several Regression-based techniques was assessed on unseen testing data and residuals were computed based on the loss functions i.e., Root Mean Squared error (RMSE), Mean Absolute Error (MAE) and Mean Squared Error (MSE), and in most cases, blended ensemble learning model resulted in least error values while computing combined effect of different abiotic and biotic factors on population fluctuations of predators. It delivered following MSE (0.0194, 3.9488, 0.3310), MAE (0.0637, 1.0549, 0.4093) and RMSE (0.1395, 1.9871, 0.5753) values while modeling Chrysopids, Coccinellids and Spiders populations respectively. Deep Neural Network also performed well in current domain of research after ensemble model.
No takes yet. Share an insight, caveat, or question.
Bibi et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: