Table 1.
Machine learning (ML) techniques for avian influenza. Abbreviations: Artificial Neural Networks (ANNs); Convolutional Neural Network (CNN); Genetic Algorithm for Rule-Set Prediction (GARP); Maximum Entropy (MaxEnt); Species Distribution Modeling (SDM); Single Shot MultiBox Detector (SSD); Support Vector Machine (SVM); Extreme Gradient Boosting (XGBoost); You Only Look Once (YOLO).
| ML Method | Application | Animal Health | Human Health |
|---|---|---|---|
| Logistic regression [38,39,40,41,42,43,44,45,46,48,51,52], tobit regression [47], negative binomial regression [49], linear regression [52] | Identify animal and environmental risk factors associated with avian influenza occurrence | √ | |
| Logistic regression [55,56] | Identify risk factors that result in the transmission of avian influenza from birds to mammalians such as dogs, cats, and pigs | √ | |
| Logistic regression [45], Poisson regression [50], multivariable regression [53], linear regression [57] | Identify environmental, demographic, and socioeconomic risk factors associated with avian influenza occurrence | √ | |
| Linear regression [58], multilevel regression [59], birth process with regression model [60], logistic regression [61], SVM [62] | Study the efficiency of preventive policies such as poultry vaccination on the spread of the avian influenza virus among birds | √ | |
| Cox proportional hazards regression [32], logistic regression [63,64] | Study the efficiency of pharmaceutical and non-pharmaceutical interventions on avian influenza transmission and mortality | √ | |
| Gradient boosted tree [65], SVM [66], multiple linear regression [67], simple regression [68], logistic regression [39,69,70,71] | Identify the molecular signatures that define the pathogenicity of viral strains | √ | |
| Deep CNN [72], logistic regression [73] | Predict genomic sequences | √ | |
| Random Forest, Gradient Boosting, and XGBoost [74], SVM and ANN [75], binomial regression [76], and deep-learning models [77,78] | Predict avian influenza outbreaks in animals at the temporal level | √ | |
| Multiple linear regression [79] | Forecast avian influenza outbreaks in humans at the temporal level | √ | |
| Bayesian logistic regression, XGBoost [41,80,81], spatial regression analysis [41,82], region-based CNN, SSD and YOLO [83], logistic regression [84,85], generalized linear mixed model [86], Poisson and logistic regression [87] | Identify geographical regions and risk factors of avian influenza hotspots | √ | |
| MaxEnt [88,89,90], GARP [91], Random Forest [90] | Identify geographical and spatial factors of migratory bird hotspots and provide a risk map using SDM | √ | |
| Linear regression and spatial regression [82], logistic regression [92,93,94,95], boosted regression tree [96], Poisson regression [97] | Analyze spatiotemporal factors affecting avian influenza | √ |