Predicting insect outbreaks using machine learning: A mountain pine beetle case study

Ramazi, Pouria; Kunegel-Lion, Mélodie; Greiner, Russell; Lewis, Mark A.

doi:10.1002/ece3.7921

Cited by 26 publications

(7 citation statements)

References 67 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Additionally, we assumed no manipulation, or forced structure in construction of the BN, as to allow the entire structure to be learned from the training data set only. We then learned the BN parameters, and compared the prediction accuracy of the resulting BN with four predictive models: generalized linear model (GLM), naive Bayes (NB), boosted decision tree (or, gradient boosting machine (GBM)), k nearest neighbor (KNN), and neural network (NN) (Ramazi et al., 2021b). We used the area under the curve (AUC) (Hajian‐Tilaki, 2013) as the performance measure over the test data set.…”

Section: Methodsmentioning

confidence: 99%

Predicting Imminent Cyanobacterial Blooms in Lakes Using Incomplete Timely Data

Heggerud,

Xu,

Wang

et al. 2024

Water Resources Research

Self Cite

View full text Add to dashboard Cite

Toxic cyanobacterial blooms (CBs) are becoming more frequent globally, posing a threat to freshwater ecosystems. While making long‐range forecasts is overly challenging, predicting imminent CBs is possible from precise monitoring data of the underlying covariates. It is, however, infeasibly costly to conduct precise monitoring on a large scale, leaving most lakes unmonitored or only partially monitored. The challenge is hence to build a predictive model that can use the incomplete, partially‐monitored data to make near‐future CB predictions. By using 30 years of monitoring data for 78 water bodies in Alberta, Canada, combined with data of watershed characteristics (including natural land cover and anthropogenic land use) and meteorological conditions, we train a Bayesian network that predicts future 2‐week CB with an area under the curve (AUC) of 0.83. The only monitoring data that the model needs to reach this level of accuracy are whether the cell count and Secchi depth are low, medium, or high, which can be estimated by advanced high‐resolution imaging technology or trained local citizens. The model is robust against missing values as in the absence of any single covariate, it performs with an AUC of at least 0.78. While taking a major step toward reduced‐cost, less data‐intensive CB forecasting, our results identify those key covariates that are worth the monitoring investment for highly accurate predictions.

show abstract

Section: Methodsmentioning

confidence: 99%

Predicting Imminent Cyanobacterial Blooms in Lakes Using Incomplete Timely Data

Heggerud,

Xu,

Wang

et al. 2024

Water Resources Research

Self Cite

View full text Add to dashboard Cite

show abstract

“…As this is a time-series prediction task, the division was done temporally , where the beginning “chunk” of the data is taken as the training and the remaining chunk as the test dataset (Ramazi et al. 2021b , a ). Moreover, to increase evaluation reliability, several training and test durations were considered.…”

Section: Methodsmentioning

confidence: 99%

From Policy to Prediction: Forecasting COVID-19 Dynamics Under Imperfect Vaccination

Wang

Ramazi

et al. 2022

Bull Math Biol

Self Cite

View full text Add to dashboard Cite

Understanding the joint impact of vaccination and non-pharmaceutical interventions on COVID-19 development is important for making public health decisions that control the pandemic. Recently, we created a method in forecasting the daily number of confirmed cases of infectious diseases by combining a mechanistic ordinary differential equation (ODE) model for infectious classes and a generalized boosting machine learning model (GBM) for predicting how public health policies and mobility data affect the transmission rate in the ODE model (Wang et al. in Bull Math Biol 84:57, 2022). In this paper, we extend the method to the post-vaccination period, accordingly obtain a retrospective forecast of COVID-19 daily confirmed cases in the US, and identify the relative influence of the policies used as the predictor variables. In particular, our ODE model contains both partially and fully vaccinated compartments and accounts for the breakthrough cases, that is, vaccinated individuals can still get infected. Our results indicate that the inclusion of data on non-pharmaceutical interventions can significantly improve the accuracy of the predictions. With the use of policy data, the model predicts the number of daily infected cases up to 35 days in the future, with an average mean absolute percentage error of 20.15%, which is further improved to 14.88% if combined with human mobility data. Moreover, the most influential predictor variables are the policies of restrictions on gatherings, testing B Hao Wang

show abstract

“…To validate the submodel for angler traffic between localities and subbasins, we randomly split the app data into a training (fitting) and a testing (validation) dataset, each containing observations for half of the anglers respectively. Note that a random split is in line with our purpose of evaluating the model accuracy in predicting unreported trips, and hence, a temporal split used for evaluating the model accuracy in making future predictions is not needed (Ramazi, Kunegel-Lion, et al, 2021). We fitted our model to the training data and computed the mean yearly number of recorded angler trips for each origin, destination and origin-destination pair.…”

Section: Model Evaluationmentioning

confidence: 99%

Boosting propagule transport models with individual‐specific data from mobile apps

Fischer

Ramazi

Simmons³

et al. 2023

Journal of Applied Ecology

Self Cite

View full text Add to dashboard Cite

Management of invasive species and pathogens requires information about the traffic of potential vectors. Such information is often taken from vector traffic models fitted to survey data. Here, user‐specific data collected via mobile apps offer new opportunities to obtain more accurate estimates and to analyse how vectors' individual preferences affect propagule flows. However, data voluntarily reported via apps may lack some trip records, adding a significant layer of uncertainty. We show how the benefits of app‐based data can be exploited despite this drawback. Based on data collected via an angler app, we built a stochastic model for angler traffic in the Canadian province Alberta. There, anglers facilitate the spread of whirling disease, a parasite‐induced fish disease. The model is temporally and spatially explicit and accounts for individual preferences and repeating behaviour of anglers, helping to address the problem of missing trip records. We obtained estimates of angler traffic between all subbasins in Alberta. The model's accuracy exceeds that of direct empirical estimates even when fewer data were used to fit the model. The results indicate that anglers' local preferences and their tendency to revisit previous destinations reduce the number of long interwaterbody trips potentially dispersing whirling disease. According to our model, anglers revisit their previous destination in 64% of their trips, making these trips irrelevant for the spread of whirling disease. Furthermore, 54% of fishing trips end in individual‐specific spatially contained areas with mean radius of 54.7 km. Finally, although the fraction of trips that anglers report was unknown, we were able to estimate the total yearly number of fishing trips in Alberta, matching an independent empirical estimate. Policy implications. We make two major contributions: (1) we provide a model that uses mobile app data to boost the mechanistic accuracy of classic propagule transport models, and (2) we demonstrate the importance of individual‐specific behaviour of vectors for propagule transport. Ignoring vectors' local preferences and their tendency to revisit previous destinations can lead to significant overestimates of vector traffic and biased estimates of propagule flows. This has clear implications for the management of invasive species and animal diseases.

show abstract

Predicting insect outbreaks using machine learning: A mountain pine beetle case study

Abstract: This is an open access article under the terms of the Creat ive Commo ns Attri bution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Cited by 26 publications

References 67 publications

Predicting Imminent Cyanobacterial Blooms in Lakes Using Incomplete Timely Data

Predicting Imminent Cyanobacterial Blooms in Lakes Using Incomplete Timely Data

From Policy to Prediction: Forecasting COVID-19 Dynamics Under Imperfect Vaccination

Boosting propagule transport models with individual‐specific data from mobile apps

Contact Info

Product

Resources

About