Direct Multi‐step Estimation and Forecasting

Chevillon, Guillaume

doi:10.1111/j.1467-6419.2007.00518.x

Cited by 140 publications

(53 citation statements)

References 65 publications

(117 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Along this line, the identification of multi-step-ahead optimal predictors of certain observed, i.e., available to measurements, variables have been analyzed in the past by several authors (e.g., [Åström, 1980], [Anderson and Moore, 1979], [Holst, 1977], [Mosca et al, 1989], [Weiss, 1991], [Chevillon, 2007]). This chapter will be focused on the direct identification of multistep-ahead predictors and their application to the problem of blood glucose prediction in diabetic patients.…”

Section: Linear Prediction Strategiesmentioning

confidence: 99%

Linear Modeling and Prediction in Diabetes Physiology

Cescon

2014

Lecture Notes in Bioengineering

View full text Add to dashboard Cite

Diabetes Mellitus is a chronic disease characterized by the inability of the organism to autonomously regulate the blood glucose level due to insulin deficiency or resistance, leading to serious health damages. The therapy is essentially based on insulin injections and depends strongly on patient daily decisions, being mainly based upon empirical experience and rules of thumb. The development of a prediction engine capable of personalized on-the-spot decision making concerning the most adequate choice of insulin delivery, meal intake and exercise would therefore be a valuable initiative towards an improved management of the desease.This thesis presents work on data-driven glucose metabolism modeling and short-term, that is, up to 120 minutes, blood-glucose prediction in Type 1 Diabetes Mellitus (T1DM) subjects.In order to address model-based control for blood glucose regulation, low-order, individualized, data-driven, stable, physiological relevant models were identified from a population of 9 T1DM patients data. Model structures include: autoregressive moving average with exogenous inputs (ARMAX) models and state-space models.ARMAX multi-step-ahead predictors were estimated by means of least-squares estimation; next regularization of the autoregressive coefficients was introduced. ARMAX-based predictors and zero-order hold were computed to allow comparison.Finally, preliminary results on subspace-based multi-step-ahead multivariate predictors is presented.

show abstract

Section: Linear Prediction Strategiesmentioning

confidence: 99%

Linear Modeling and Prediction in Diabetes Physiology

Cescon

2014

Lecture Notes in Bioengineering

View full text Add to dashboard Cite

show abstract

“…The data for these time series are generated by the Chaotic The NN3 competition was organized in 2007, targeting computational-intelligence forecasting approaches. The competition dataset of 111 monthly time series drawn from homogeneous population of real business time series is used for evaluation 2 As such, three datasets of 20 Logistic time series, 20 Mackey-Glass time series, and 111 NN3 time series are used for evaluating the performances of the proposed PSO-MISMO and the counterparts in this study. Each series is split into an estimation sample and a hold-out sample.…”

Section: A Datasets Descriptionmentioning

confidence: 99%

PSO-MISMO Modeling Strategy for MultiStep-Ahead Time Series Prediction

Bao

Xiong

2014

IEEE Trans. Cybern.

View full text Add to dashboard Cite

Abstract-Multi-step-ahead time series prediction is one of the most challenging research topics in the field of time series modeling and prediction, and is continually under research. Recently, the multiple-input several multiple-outputs (MISMO) modeling strategy has been proposed as a promising alternative for multi-step-ahead time series prediction, exhibiting advantages compared with the two currently dominating strategies, the iterated and the direct strategies. Built on the established MISMO strategy, this study proposes a particle swarm optimization (PSO)-based MISMO modeling strategy, which is capable of determining the number of sub-models in a self-adaptive mode, with varying prediction horizons. Rather than deriving crisp divides with equal-size s prediction horizons from the established MISMO, the proposed PSO-MISMO strategy, implemented with neural networks, employs a heuristic to create flexible divides with varying sizes of prediction horizons and to generate corresponding sub-models, providing considerable flexibility in model construction, which has been validated with simulated and real datasets.

show abstract

“…Previsão multipassos à frente de séries temporais é um grande desafio em diversas áreas do conhecimento tais como economia (Chevillon, 2007, Marcellino et al, 2005 e hidrologia (Bhagwat et al, 2012, Pionosi e Raso, 2012, Bao et al,2012, pois quanto maior o horizonte de previsão, maior é a incerteza (Ben Taieb et al, 2011).…”

Section: Introductionunclassified

Análise e comparação de estratégias de previsão multipassos à frente para vazão afluente em bacias hidrográficas

Costa¹,

Mendes²

2013

Proceeding Series of the Brazilian Society of Computational and Applied Mathematics

View full text Add to dashboard Cite

AbstractOne of the problems in the time series literature is to obtain multistep ahead predictions with good accuracy. In order to achieve this goal, not only the algorithm used but the strategy applied is of great importance. In this work, to improve the performance of the estimation algorithm, five different strategies for MultiStep ahead predictions, Direct, Recursive, DirRec, MIMO and DIRMO are compared and analyzed in the context of a real prediction problem. The piece of real data used here contains precipitation and riverflow measurements of a small hydroelectric power station in the state of Minas Gerais, UHE Rosal Cemig. The results show that the Recursive strategy performed better while the strategy DirRec exhibits the best MAPE index.Keywords Prediction, Multi-step, Time-SeriesResumoA obtenção de uma boa Predição de Multipassos à frente é um dos grandes objetivos em diversas áreas da ciência que lidam com com séries temporais. Para alcançar este objetivo, não só o algoritmo utilizado bem como a estratégia aplicada para a determinação da predição devem ser considerados. Neste trabalho, visando obter o melhor desempenho do algoritmo utilizado no difícil problema de predição de séries reais de vazão, discute-se cinco estratégias de predição multipassos à frente hoje disponíveis na literatura técnica: Direta, Recursiva, DirRec, MIMO e DIRMO. Os dados utilizados, precipitação e vazão, são reais e referem-se a uma bacia hidrográfica de pequeno porte. Os resultados obtidos apontam que a estratégia Recursiva teve melhor desempenho mas a estratégia DirRec obteve o melhor índice MAPE.

show abstract

Direct Multi‐step Estimation and Forecasting

Cited by 140 publications

References 65 publications

Linear Modeling and Prediction in Diabetes Physiology

Linear Modeling and Prediction in Diabetes Physiology

PSO-MISMO Modeling Strategy for MultiStep-Ahead Time Series Prediction

Análise e comparação de estratégias de previsão multipassos à frente para vazão afluente em bacias hidrográficas

Contact Info

Product

Resources

About