This study explored the usefulness and implications of the Bayesian hyperparameter optimization in developing species distribution models (SDMs). A variety of machine learning (ML) algorithms, namely, support vector machine (SVM), random forest (RF), boosted regression tree (BRT), XGBoost (XGB), and Multilayer perceptron (MLP) were used for predicting the occurrence of four benthic macroinvertebrate species. The Bayesian optimization method successfully tuned model hyperparameters, with all ML models resulting an area under the curve (AUC) > 0.7. Also, hyperparameter search ranges that generally clustered around the optimal values suggest the efficiency of the Bayesian optimization in finding optimal sets of hyperparameters. Tree based ensemble algorithms (BRT, RF, and XGB) tended to show higher performances than SVM and MLP. Important hyperparameters and optimal values differed by species and ML model, indicating the necessity of hyperparameter tuning for improving individual model performances. The optimization results demonstrate that for all macroinvertebrate species SVM and RF required fewer numbers of trials until obtaining optimal hyperparameter sets, leading to reduced computational cost compared to other ML algorithms. The results of this study suggest that the Bayesian optimization is an efficient method for hyperparameter optimization of machine learning algorithms.

In the present study, a novel ELISA method used recombinant nucleocapsid protein (rNP) as the coating agent. Recombinant Newcastle disease virus (NDV) protein was cloned and expressed in Escherichia coli. Though the rNP-ELISA results were consistent with commercial ELISA results for the NDV-negative sera samples, qualitatively and quantitatively variable (often reduced) results were obtained with NDV-positive sera. Although the rNP-ELISA results for NDV detection were inconclusive, further improvement and standardization of the rNP-ELISA approach, such as using multiple recombinant proteins as the ELISA coating agent and performing comprehensive statistical analyses of combined recombinant protein ELISA, should help counter Newcastle disease outbreaks by improving NDV detection.

To understand the genesis of tropical cyclones (TC), we computed TC genesis probability (GPr) by partitioning a highly localized genesis frequency (GFq) into nearby grid boxes in proportion to the spatial coherence of genesis potential index (GPI). From the analysis of TCs simulated by the Seoul National University Atmosphere Model Version 0 and the observed TCs, it was shown that GPr reasonably converges to GFq when averaged over a long-term period in a decent grid size, supporting its validity as a proxy representing a true TC GPr. The composite anomalies of the gridded GPr in association with the Asia summer monsoon, El Nino-Southern Oscillation (ENSO), and the Madden-Julian Oscillation (MJO) are much less noisy than those of GFq, and consequently are better interpretable. In summary, GPr converges to GFq, varies more smoothly than GFq, represents the spatiotemporal variations of GFq better than GPI, and depicts GFq with greater spatial details than other spatially smoothed GFqs.