Prediction of the Spatial Distribution of Bovine Endemic Fluorosis Using Ordinary Kriging

Abstract

The aim of the studies was to develop an alternative method which could overcome the lack of sampling to improve the efficiency of control efforts for bovine endemic fluorosis. The spatial distribution characteristics of the disease were analysed and a prediction model for the estimation of fluorosis distribution in some districts in northwest Liaoning province in China was established. The model used ordinary kriging, and was evaluated using cross-validation. Analysis showed that the distribution of the disease was spatial autocorrelation. The prediction error of the cross-validation (ME = -0.0092, PMSE = 0.627, AKSE = 0.597, and RMSP = 1.007) and comparison with the actual disease distribution indicated that the prediction map accurately distributed bovine endemic fluorosis. It is feasible to predict bovine endemic fluorosis in the area by using ordinary kriging and limited data.

Application and Comparison of Metaheuristic and New Metamodel Based Global Optimization Methods to the Optimal Operation of Active Distribution Networks

“…Another type of effective solution method is the surrogate or metamodel-based global optimization (MBGO) search techniques, which have been introduced and investigated by many researchers, including the authors' group [21]. The approach uses limited "expensive" sample data points from the original, computationally expensive optimization model to introduce a surrogate model, or metamodel, such as Kriging [22] and radial basis functions (RBF) [23], and to effectively use the "cheaper" sample points from the metamodel to speed up the search of the global optimum with much reduced number of original model evaluations and computational time. Several reviews have systematically presented the advantages of these algorithms [21,[24][25][26].…”