Prediction Of Hydrological Models’ Uncertainty By A Committee Of Machine Learning-Models

This study presents an approach to combine uncertainties of the hydrological model outputs predicted from a number of machine learning models. The machine learning based uncertainty prediction approach is very useful for estimation of hydrological models' uncertainty in particular hydro-metrological situation in real-time application [1]. In this approach the hydrological model realizations from Monte Carlo simulations are used to build different machine learning uncertainty models to predict uncertainty (quantiles of pdf) of the a deterministic output from hydrological model . Uncertainty models are trained using antecedent precipitation and streamflows as inputs. The trained models are then employed to predict the model output uncertainty which is specific for the new input data. We used three machine learning models namely artificial neural networks, model tree, locally weighted regression to predict output uncertainties. These three models produce similar verification results, which can be improved by merging their outputs dynamically. We propose an approach to form a committee of the three models to combine their outputs. The approach is applied to estimate uncertainty of streamflows simulation from a conceptual hydrological model in the Brue catchment in UK and the Bagmati catchment in Nepal. The verification results show that merged output is better than an individual model output. [1] D. L. Shrestha, N. Kayastha, and D. P. Solomatine, and R. Price. Encapsulation of parameteric uncertainty statistics by various predictive machine learning models: MLUE method, Journal of Hydroinformatic, in press, 2013.

A Gradient-Type Method For Real-Time State Estimation Of Water Distribution Networks

Drinking water distribution networks risk exposure to malicious or accidental contamination. Several levels of responses are conceivable. One of them consists to install a sensor network to monitor the system on real time. Once a contamination has been detected, this is also important to take appropriate counter-measures. In the SMaRT-OnlineWDN project, this relies on modeling to predict both hydraulics and water quality. An online model use makes identification of the contaminant source and simulation of the contaminated area possible. The objective of this paper is to present SMaRT-OnlineWDN experience and research results for hydraulic state estimation with sampling frequency of few minutes. A least squares problem with bound constraints is formulated to adjust demand class coefficient to best fit the observed values at a given time. The criterion is a Huber function to limit the influence of outliers. A Tikhonov regularization is introduced for consideration of prior information on the parameter vector. Then the Levenberg-Marquardt algorithm is applied that use derivative information for limiting the number of iterations. Confidence intervals for the state prediction are also given. The results are presented and discussed on real networks in France and Germany.

Model Prediction Control For Water Management Using Adaptive Prediction Accuracy

In the field of operational water management, Model Predictive Control (MPC) has gained popularity owing to its versatility and flexibility. The MPC controller, which takes predictions, time delay and uncertainties into account, can be designed for multi-objective management problems and for large-scale systems. Nonetheless, a critical obstacle, which needs to be overcome in MPC, is the large computational burden when a large-scale system is considered or a long prediction horizon is involved. In order to solve this problem, we use an adaptive prediction accuracy (APA) approach that can reduce the computational burden almost by half. The proposed MPC scheme with this scheme is tested on the northern Dutch water system, which comprises Lake IJssel, Lake Marker, the River IJssel and the North Sea Canal. The simulation results show that by using the MPC-APA scheme, the computational time can be reduced to a large extent and a flood protection problem over longer prediction horizons can be well solved.

Previsão em tempo atual de cheias com uso de sistema especialista difuso

Sistemas de previsão de cheias podem ser adequadamente utilizados quando o alcance é suficiente, em comparação com o tempo necessário para ações preventivas ou corretivas. Além disso, são fundamentalmente importantes a confiabilidade e a precisão das previsões. Previsões de níveis de inundação são sempre aproximações, e intervalos de confiança não são sempre aplicáveis, especialmente com graus de incerteza altos, o que produz intervalos de confiança muito grandes. Estes intervalos são problemáticos, em presença de níveis fluviais muito altos ou muito baixos. Neste estudo, previsões de níveis de cheia são efetuadas, tanto na forma numérica tradicional quanto na forma de categorias, para as quais utiliza-se um sistema especialista baseado em regras e inferências difusas. Metodologias e procedimentos computacionais para aprendizado, simulação e consulta são idealizados, e então desenvolvidos sob forma de um aplicativo (SELF – Sistema Especialista com uso de Lógica “Fuzzy”), com objetivo de pesquisa e operação. As comparações, com base nos aspectos de utilização para a previsão, de sistemas especialistas difusos e modelos empíricos lineares, revelam forte analogia, apesar das diferenças teóricas fundamentais existentes. As metodologias são aplicadas para previsão na bacia do rio Camaquã (15543 km2), para alcances entre 10 e 48 horas. Dificuldades práticas à aplicação são identificadas, resultando em soluções as quais constituem-se em avanços do conhecimento e da técnica. Previsões, tanto na forma numérica quanto categorizada são executadas com sucesso, com uso dos novos recursos. As avaliações e comparações das previsões são feitas utilizandose um novo grupo de estatísticas, derivadas das freqüências simultâneas de ocorrência de valores observados e preditos na mesma categoria, durante a simulação. Os efeitos da variação da densidade da rede são analisados, verificando-se que sistemas de previsão pluvio-hidrométrica em tempo atual são possíveis, mesmo com pequeno número de postos de aquisição de dados de chuva, para previsões sob forma de categorias difusas.

The combination of neural estimates in prediction and decision problems

In this dissertation, different ways of combining neural predictive models or neural-based forecasts are discussed. The proposed approaches consider mostly Gaussian radial basis function networks, which can be efficiently identified and estimated through recursive/adaptive methods. Two different ways of combining are explored to get a final estimate – model mixing and model synthesis –, with the aim of obtaining improvements both in terms of efficiency and effectiveness. In the context of model mixing, the usual framework for linearly combining estimates from different models is extended, to deal with the case where the forecast errors from those models are correlated. In the context of model synthesis, and to address the problems raised by heavily nonstationary time series, we propose hybrid dynamic models for more advanced time series forecasting, composed of a dynamic trend regressive model (or, even, a dynamic harmonic regressive model), and a Gaussian radial basis function network. Additionally, using the model mixing procedure, two approaches for decision-making from forecasting models are discussed and compared: either inferring decisions from combined predictive estimates, or combining prescriptive solutions derived from different forecasting models. Finally, the application of some of the models and methods proposed previously is illustrated with two case studies, based on time series from finance and from tourism.

Leaf area prediction models for Zinnia elegans Jacq., Zinnia haageana regel and 'profusion cherry'

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Haematological and biochemical reference intervals for farmed channel catfish

The reference intervals for biochemical variables and red blood cell indices of healthy intensively bred channel catfish Ictalurus punctatus were determined. The blood variables were determined using standardized clinical methods. The reference intervals (25th and 75th percentiles) were established using a non-parametric method. Reference intervals for plasma glucose, serum total protein, sodium, potassium, calcium, magnesium, chloride concentration, primary and secondary red blood cell indices were established. The haematological and biochemical reference intervals established may allow important clinical decisions about channel catfish. (c) 2007 the Authors Journal compilation (C) 2007 the Fisheries Society of the British Isles.

Cutaneous Basophilia in the Resistance of Goats to Amblyomma cajennense Nymphs after Repeated Infestations

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Guinea Pigs Develop Cutaneous Basophilia after Repeated Infestations by Nymphs of the Tick Amblyomma triste

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Latent heat loss of Holstein cows in a tropical environment: a prediction model

Nove vacas Holandesas lactantes com 526 ± 5 kg de peso corporal (cinco predominantemente pretas e quatro predominantemente brancas), criadas em região tropical e manejadas em pastagens, foram observadas com os objetivos de determinar simultaneamente as taxas de evaporação cutânea e respiratória em ambiente tropical e desenvolver modelos de predição. Para a medição da perda de calor latente pela superfície corporal, utilizou-se uma cápsula ventilada e, para a perda por respiração, utilizou-se uma máscara facial. Os resultados mostraram que as vacas que tinham maior peso corporal (classe 2 e 3) apresentaram maiores taxas evaporativas. Quando a temperatura do ar aumentou de 10 para 36ºC e a umidade relativa do ar caiu de 90 para 30%, a eliminação de calor por evaporação respiratória aumentou de aproximadamente 5 para 57 W m-2 e a evaporação na superfície corporal passou de 30 para 350 W m-2. Esses resultados confirmam que a eliminação de calor latente é o principal mecanismo de perda de energia térmica sob altas temperaturas (>30ºC); a evaporação cutânea é a maior via e corresponde a aproximadamente 85% da perda total de calor, enquanto o restante é eliminado pelo sistema respiratório. O modelo para predizer o fluxo de perda de calor latente baseado em variáveis fisiológicas e ambientais pode ser utilizado para estimar a contribuição da evaporação na termorregulação, enquanto o modelo baseado somente na temperatura do ar deve ser usado apenas para a simples caracterização do processo evaporativo.

Effects of different sampling intervals on apparent protein and energy digestibility of common feed ingredients by juvenile oscar fish (Astronotus ocellatus)

O presente estudo avaliou a digestibilidade aparente da proteína e da energia de ingredientes (farelo de soja, farinha de peixe, farelo de trigo e milho) por juvenis de apaiari (Astronotus ocellatus) usando dois diferentes intervalos de coleta (30 min. e 12h). Os 160 juvenis de apaiari utilizados (22,37 ± 3,06 g de peso corporal) foram divididos em quatro tanques rede plásticos e cilíndricos, cada um colocado em um tanque de alimentação de 1.000 L. O experimento foi inteiramente casualizado em esquema fatorial 2 x 4 (2 intervalos de coleta de fezes e 4 ingredientes foram) com quatro repetições. Os testes estatísticos não detectaram efeito da interação entre o intervalo de coleta e tipo de ingrediente nos coeficientes de digestibilidade. O intervalo de coleta não afetou a digestibilidade da proteína e da energia. As características físicas das fezes dos juvenis de apaiari aparentemente as tornam menos sensíveis à perda de nutrientes por lixiviação, permitindo intervalos de coleta maiores. A digestibilidade da proteína dos ingredientes avaliados foi semelhante, mostrando que a digestibilidade aparente de ingredientes animais e vegetais por juvenis de apaiari é eficiente. Os coeficientes de digestibilidade da energia foram maiores para a farinha de peixe e o farelo de soja comparado a farelo de trigo e milho. Ingredientes ricos em carboidratos (farelo de trigo e milho) apresentaram os piores coeficientes de digestibilidade da energia e, portanto, não são usados eficientemente pelos juvenis de apaiari.

Comparison of selective genotyping strategies for prediction of breeding values in a population undergoing selection

Genomewide marker information can improve the reliability of breeding value predictions for young selection candidates in genomic selection. However, the cost of genotyping limits its use to elite animals, and how such selective genotyping affects predictive ability of genomic selection models is an open question. We performed a simulation study to evaluate the quality of breeding value predictions for selection candidates based on different selective genotyping strategies in a population undergoing selection. The genome consisted of 10 chromosomes of 100 cM each. After 5,000 generations of random mating with a population size of 100 (50 males and 50 females), generation G(0) (reference population) was produced via a full factorial mating between the 50 males and 50 females from generation 5,000. Different levels of selection intensities (animals with the largest yield deviation value) in G(0) or random sampling (no selection) were used to produce offspring of G(0) generation (G(1)). Five genotyping strategies were used to choose 500 animals in G(0) to be genotyped: 1) Random: randomly selected animals, 2) Top: animals with largest yield deviation values, 3) Bottom: animals with lowest yield deviations values, 4) Extreme: animals with the 250 largest and the 250 lowest yield deviations values, and 5) Less Related: less genetically related animals. The number of individuals in G(0) and G(1) was fixed at 2,500 each, and different levels of heritability were considered (0.10, 0.25, and 0.50). Additionally, all 5 selective genotyping strategies (Random, Top, Bottom, Extreme, and Less Related) were applied to an indicator trait in generation G(0), and the results were evaluated for the target trait in generation G(1), with the genetic correlation between the 2 traits set to 0.50. The 5 genotyping strategies applied to individuals in G(0) (reference population) were compared in terms of their ability to predict the genetic values of the animals in G(1) (selection candidates). Lower correlations between genomic-based estimates of breeding values (GEBV) and true breeding values (TBV) were obtained when using the Bottom strategy. For Random, Extreme, and Less Related strategies, the correlation between GEBV and TBV became slightly larger as selection intensity decreased and was largest when no selection occurred. These 3 strategies were better than the Top approach. In addition, the Extreme, Random, and Less Related strategies had smaller predictive mean squared errors (PMSE) followed by the Top and Bottom methods. Overall, the Extreme genotyping strategy led to the best predictive ability of breeding values, indicating that animals with extreme yield deviations values in a reference population are the most informative when training genomic selection models.

Prediction equations to estimate the demand of energy and crude protein for maintenance, gain and egg production for laying Japanese quails

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Use of ridge regression for the prediction of early growth performance in crossbred calves

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)