Ordinal regression models for predicting deoxynivalenol in winter wheat

Landschoot, Sofie; Waegeman, Willem; Audenaert, Kris; Haesaert, Geert; Baets, Bernard De

doi:10.1111/ppa.12041

Cited by 28 publications

(18 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The overall inclusion of R-based predictors by the BRT modeling algorithm contrasts with the lack of this predictor type in logistic regression models we presented earlier (51), where R-type predictors were dropped due to low selection frequency by the algorithms used. The majority of FHB models do, in fact, include rainfall-based predictors (7,12,13,23,25,29,31,39,41,47,54). Nevertheless, the strongest predictors of FHB epidemics, based on the full brt u model, were, in general, from one of three groups: mean RH per 24-h day (group 7), mean T per day (group 16), and the number of hours (24-h day) in which T was 9 to 30°C and RH ≥ 90% simultaneously (group 36).…”

Section: Discussionmentioning

confidence: 99%

Predicting Fusarium Head Blight Epidemics with Boosted Regression Trees

Shah¹,

Wolf²,

Paul³

et al. 2014

Phytopathology®

View full text Add to dashboard Cite

Predicting major Fusarium head blight (FHB) epidemics allows for the judicious use of fungicides in suppressing disease development. Our objectives were to investigate the utility of boosted regression trees (BRTs) for predictive modeling of FHB epidemics in the United States, and to compare the predictive performances of the BRT models with those of logistic regression models we had developed previously. The data included 527 FHB observations from 15 states over 26 years. BRTs were fit to a training data set of 369 FHB observations, in which FHB epidemics were classified as either major (severity ≥ 10%) or non-major (severity < 10%), linked to a predictor matrix consisting of 350 weatherbased variables and categorical variables for wheat type (spring or winter), presence or absence of corn residue, and cultivar resistance. Predictive performance was estimated on a test (holdout) data set consisting of the remaining 158 observations. BRTs had a misclassification rate of 0.23 on the test data, which was 31% lower than the average misclassification rate over 15 logistic regression models we had presented earlier. The strongest predictors were generally one of mean daily relative humidity, mean daily temperature, and the number of hours in which the temperature was between 9 and 30°C and relative humidity ≥ 90% simultaneously. Moreover, the predicted risk of major epidemics increased substantially when mean daily relative humidity rose above 70%, which is a lower threshold than previously modeled for most plant pathosystems. BRTs led to novel insights into the weather-epidemic relationship.Additional keywords: disease modeling, disease forecasting, machine learning, plant disease epidemiology, wheat scab.Major epidemics of Fusarium head blight (FHB), caused primarily by Fusarium graminearum sensu stricto (59) of the F. graminearum species complex (43,53), are a recurring obstacle to successful wheat (Triticum aestivum L. em. Thell) production worldwide. Epidemics are responsible for large direct (35,36) and indirect (42) economic losses. Foliar fungicide applications timed to coincide with anthesis, if the environment is conducive for FHB, are one component of an effective disease management strategy (56). Accurate forecasts help growers recognize when their wheat crops are at a high risk of a major FHB epidemic, and most likely to benefit from a fungicide. When the risk of a major epidemic is low, growers could forgo inessential fungicide applications. Management decisions (and, hence, forecasts) need to be made by anthesis so that producers have sufficient time to spray should they choose that option.The Fusarium Head Blight Risk Assessment Tool (http://www. wheatscab.psu.edu) is a publicly funded service that provides local-level, empirical FHB predictions across several of the wheat-growing regions of the United States which have historically experienced FHB epidemics (35). The forecasts are based on logistic regression models developed by De Wolf et al. (13), with subsequent revisions (37,38). We recently reex...

show abstract

Section: Discussionmentioning

confidence: 99%

Predicting Fusarium Head Blight Epidemics with Boosted Regression Trees

Shah¹,

Wolf²,

Paul³

et al. 2014

Phytopathology®

View full text Add to dashboard Cite

show abstract

“…Provisional models often focus on weather forecasts and the susceptibility of the planted cultivar. Modelling mycotoxin production is more difficult with respect to foreseeing disease incidence and severity since toxin production is affected by additional factors, e.g., strain pathogenicity, toxigenicity, competition with other microbes in the plant, and effects of fungicides on toxin biosynthesis [ 68 , 69 ]. Most of the existing models are empirical in nature, as the fundamental factors connecting disease progression and toxin production to the environment are not well understood.…”

Section: Provisional Model and Historical Data In Relation To A Nementioning

confidence: 99%

Toxigenic Fungi and Mycotoxins in a Climate Change Scenario: Ecology, Genomics, Distribution, Prediction and Prevention of the Risk

et al. 2020

View full text Add to dashboard Cite

Toxigenic fungi and mycotoxins are very common in food crops, with noticeable differences in their host specificity in terms of pathogenicity and toxin contamination. In addition, such crops may be infected with mixtures of mycotoxigenic fungi, resulting in multi-mycotoxin contamination. Climate represents the key factor in driving the fungal community structure and mycotoxin contamination levels pre- and post-harvest. Thus, there is significant interest in understanding the impact of interacting climate change-related abiotic factors (especially increased temperature, elevated CO2 and extremes in water availability) on the relative risks of mycotoxin contamination and impacts on food safety and security. We have thus examined the available information from the last decade on relative risks of mycotoxin contamination under future climate change scenarios and identified the gaps in knowledge. This has included the available scientific information on the ecology, genomics, distribution of toxigenic fungi and intervention strategies for mycotoxin control worldwide. In addition, some suggestions for prediction and prevention of mycotoxin risks are summarized together with future perspectives and research needs for a better understanding of the impacts of climate change scenarios.

show abstract

“…Our models were also restricted to wheat production in the U.S. and even in that, do not cover the western States where FHB is much less common [ 41 ] and where field observations were not available. Other responses have been modeled in the FHB system, including grain contamination with the mycotoxins deoxynivalenol and zearalenone at harvest [ 40 , 42 ], indices of disease level or of mycotoxin concentration [ 34 , 43 ], ordinal representations of disease levels [ 44 ], and disease incidence directly [ 45 ]. These responses are clearly on different scales and represent different disease aspects (symptoms or toxin concentration, for example), and therefore ensembling across these models would be more challenging unless their disparate responses were somehow expressed in a common unit.…”

Section: Discussionmentioning

confidence: 99%

Accuracy in the prediction of disease epidemics when ensembling simple but highly correlated models

et al. 2021

View full text Add to dashboard Cite

Ensembling combines the predictions made by individual component base models with the goal of achieving a predictive accuracy that is better than that of any one of the constituent member models. Diversity among the base models in terms of predictions is a crucial criterion in ensembling. However, there are practical instances when the available base models produce highly correlated predictions, because they may have been developed within the same research group or may have been built from the same underlying algorithm. We investigated, via a case study on Fusarium head blight (FHB) on wheat in the U.S., whether ensembles of simple yet highly correlated models for predicting the risk of FHB epidemics, all generated from logistic regression, provided any benefit to predictive performance, despite relatively low levels of base model diversity. Three ensembling methods were explored: soft voting, weighted averaging of smaller subsets of the base models, and penalized regression as a stacking algorithm. Soft voting and weighted model averages were generally better at classification than the base models, though not universally so. The performances of stacked regressions were superior to those of the other two ensembling methods we analyzed in this study. Ensembling simple yet correlated models is computationally feasible and is therefore worth pursuing for models of epidemic risk.

show abstract

Ordinal regression models for predicting deoxynivalenol in winter wheat

Cited by 28 publications

References 36 publications

Predicting Fusarium Head Blight Epidemics with Boosted Regression Trees

Predicting Fusarium Head Blight Epidemics with Boosted Regression Trees

Toxigenic Fungi and Mycotoxins in a Climate Change Scenario: Ecology, Genomics, Distribution, Prediction and Prevention of the Risk

Accuracy in the prediction of disease epidemics when ensembling simple but highly correlated models

Contact Info

Product

Resources

About