Calibration and verification of risk algorithms using logistic regression

Yuen, Jonathan; Twengström, Eva; Sigvald, Roland

doi:10.1007/bf01877054

Cited by 85 publications

(75 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Contingency tables of independent variables were constructed to represent the bivariate distribution of disease levels according to classifications of year, cropping systems and irrigation ( Table 1). The association of early blight incidence, lesion numbers and disease severity with independent variables was analyzed using logistic regression as described (Yuen et al 1996). The logistic regression model was used to evaluate the importance of multiple independent variables (cropping systems, irrigation, and year) that affect the response variable (early blight incidence, severity and lesion numbers) by calculating the probability of a given binary outcome (response) as a function of the independent variables (MacCullagh and Nelder 1989).…”

Section: Discussionmentioning

confidence: 99%

The effect of cropping systems and irrigation management on development of potato early blight

et al. 2009

View full text Add to dashboard Cite

Section: Discussionmentioning

confidence: 99%

The effect of cropping systems and irrigation management on development of potato early blight

et al. 2009

View full text Add to dashboard Cite

“…The key contributions in plant pathology have been by J. Yuen, G. Hughes, and some of their colleagues (Yuen et al, 1996;Hughes et al, 1999;Yuen and Hughes, 2002;Yuen, 2003). A very recent and thorough example is Turechek and Wilcox (2005).…”

Section: Decision Making In Epidemiologymentioning

confidence: 96%

Botanical Epidemiology: Some Key Advances and its Continuing Role in Disease Management

Madden

2006

Eur J Plant Pathol

122

View full text Add to dashboard Cite

Epidemiology involves the study of the temporal, spatial, and spatio-temporal dynamics of disease in populations, and the utilization of results of experiments and surveys to describe, understand, compare, and predict epidemics. Such understanding and description of epidemics can lead directly to the development and evaluation of efficient control strategies and tactics. Mathematical and statistical models are key tools of the epidemiologist. Recent advances in statistics, including linear and nonlinear mixed models, are allowing a more appropriate matching of data type and experimental (or survey) design to the statistical model used for analysis, in order to meet the objectives of the investigator. Coupled ordinary and partial differential equations, as well as simpler growth-curve equations, are especially useful deterministic models for representing plant disease development in fields in time and space over single seasons or many years, and their use can lead to appraisal of control strategies through metrics such as the basic reproduction number, a summary parameter that may be calculated for many general epidemic scenarios. Recently, compelling arguments have been made for the use of Bayesian decision theory in developing and evaluating real-time disease prediction rules, based on measured disease or weather conditions and either empirical or mechanistic models for disease or control intervention. Through some simple calculations of predictor accuracy and (prior) probability of an epidemic (or the need for control), the success of any predictor can be quantified in terms of the estimated probability of random observations being epidemics when predicted to be epidemics or not epidemics. Overall, despite the many contributions in epidemiology over the past four decades, more effort is still needed to convince those outside of epidemiology to more fully use epidemiological results and insights into the development and evaluation of disease controls.

show abstract

“…For technical details of generalized linear model (GLM) methodology, readers are referred to Collett (2003). For examples of its application in the development of a crop protection decision tool see Yuen et al (1996) and Burnett and Hughes (2004).…”

Section: Analytical Approachesmentioning

confidence: 99%