scholarly journals Forecasting intermittent and sparse time series: A unified probabilistic framework via deep renewal processes

PLoS ONE ◽  
2021 ◽  
Vol 16 (11) ◽  
pp. e0259764
Author(s):  
Ali Caner Türkmen ◽  
Tim Januschowski ◽  
Yuyang Wang ◽  
Ali Taylan Cemgil
Keyword(s):  
Neural Network ◽  
Time Series ◽  
Discrete Time ◽  
Predictive Accuracy ◽  
Demand Forecasting ◽  
Trivial Extension ◽  
Renewal Processes ◽  
Probabilistic Forecasting ◽  
Unified Framework ◽  
Intermittent Demand

Intermittency are a common and challenging problem in demand forecasting. We introduce a new, unified framework for building probabilistic forecasting models for intermittent demand time series, which incorporates and allows to generalize existing methods in several directions. Our framework is based on extensions of well-established model-based methods to discrete-time renewal processes, which can parsimoniously account for patterns such as aging, clustering and quasi-periodicity in demand arrivals. The connection to discrete-time renewal processes allows not only for a principled extension of Croston-type models, but additionally for a natural inclusion of neural network based models—by replacing exponential smoothing with a recurrent neural network. We also demonstrate that modeling continuous-time demand arrivals, i.e., with a temporal point process, is possible via a trivial extension of our framework. This leads to more flexible modeling in scenarios where data of individual purchase orders are directly available with granular timestamps. Complementing this theoretical advancement, we demonstrate the efficacy of our framework for forecasting practice via an extensive empirical study on standard intermittent demand data sets, in which we report predictive accuracy in a variety of scenarios.

scholarly journals Intermittent demand forecasting for medical consumables with short life cycle using a dynamic neural network during the COVID-19 epidemic

2020 ◽  
Vol 26 (4) ◽  
pp. 3106-3122
Author(s):  
Peipei Liu
Keyword(s):  
Neural Network ◽  
Life Cycle ◽  
Model Selection ◽  
Demand Forecasting ◽  
Seasonal Adjustment ◽  
Intermittent Demand ◽  
Short Life ◽  
Dynamic Neural Network ◽  
Demand Models ◽  
The Impact

Accurate demand forecasting is always critical to supply chain management. However, many uncertain factors in the market make this issue a huge challenge. Especially during the current COVID-19 outbreak, the shortage of certain types of medical consumables has become a global problem. The intermittent demand forecast of medical consumables with a short life cycle brings some new challenges, such as the demand occurring randomly in many time periods with zero demand. In this research, a seasonal adjustment method is introduced to deal with seasonal influences, and a dynamic neural network model with optimized model selection procedure and an appropriate model selection criterion are introduced as the main forecasting models. In addition, in order to reduce the impact of zero demand, it adds some input nodes to the neural network by preprocessing the original input data. Lastly, a modified error measurement method is proposed for performance evaluation. Experimental results show that the proposed forecasting framework is superior to other intermittent demand models.

Time series formed from the superposition of discrete renewal processes

1989 ◽  
Vol 26 (01) ◽  
pp. 189-195
Author(s):  
P. A. Blight
Keyword(s):  
Time Series ◽  
Discrete Time ◽  
Conditional Distribution ◽  
Counting Process ◽  
Correlation Structure ◽  
Renewal Processes ◽  
Discrete Time Series

The superposition of independent, discrete, renewal processes produces a counting process which is also a discrete time series. The conditional distribution and correlation structure of this kind of time series may be obtained. In suitable conditions the conditional distribution has a spectrum which is exactly or approximately rational. When this is so, an ARMA can be found which matches the spectrum of the superposition.

Time series formed from the superposition of discrete renewal processes

1989 ◽  
Vol 26 (1) ◽  
pp. 189-195 ◽  
Author(s):  
P. A. Blight
Keyword(s):  
Time Series ◽  
Discrete Time ◽  
Conditional Distribution ◽  
Counting Process ◽  
Correlation Structure ◽  
Renewal Processes ◽  
Discrete Time Series

The superposition of independent, discrete, renewal processes produces a counting process which is also a discrete time series. The conditional distribution and correlation structure of this kind of time series may be obtained. In suitable conditions the conditional distribution has a spectrum which is exactly or approximately rational. When this is so, an ARMA can be found which matches the spectrum of the superposition.

scholarly journals AR (p) para previsão de demanda em uma agroindústria

ForScience ◽  
2020 ◽  
Vol 8 (1) ◽  
pp. e00507
Author(s):  
Álvaro de Aquino Santos ◽  
Marcos Antônio Alves ◽  
Elias Silva De Medeiros ◽  
Igor Neves Nunes ◽  
Rafael João de Melo Miguel Cardoso
Keyword(s):  
Time Series ◽  
Predictive Accuracy ◽  
Demand Forecasting ◽  
Evaluation Criteria ◽  
Parametric Model ◽  
Minas Gerais ◽  
Broiler Chicks ◽  
First Order ◽  
Satisfactory Performance ◽  
Historical Series

Os modelos de previsão de séries temporais são ferramentas importantes no apoio ao planejamento da produção e na tomada de decisão de organizações como as agroindústrias. Um dos desafios da agroindústria brasileira de frango, uma das mais importantes do mundo, é prever adequadamente a demanda de seus clientes. Nesse contexto, o objetivo deste estudo foi propor um modelo paramétrico para previsão de demanda baseado nos dados de expedição de pintos da linhagem de corte em uma agroindústria do centro-oeste de Minas Gerais. Diferentes métodos foram avaliados sobre a série histórica de 59 semanas, a fim de identificar o comportamento das expedições semanais e verificar possíveis tendências e sazonalidades. Dados de 56 semanas foram avaliados e os modelos candidatos foram obtidos a partir da análise das autocorrelações das observações. Por meio dos critérios de avaliação BIC e AIC, o modelo Autorregressivo de Primeira Ordem (AR) (1) se mostrou o mais adequado. Para avaliar o poder de predição do modelo AR (1) foi realizada uma comparação entre os valores preditos e observados nas últimas quatro semanas da série. Por meio da análise foi verificado um desempenho satisfatório, uma vez que os valores observados se encontravam dentro do intervalo de 95% de confiança construído por meio do modelo. Palavras-chave: Série temporal. Previsão de demanda. Pintos de corte. Frango. AR (p) for forecasting demanddata in an agroindustry  Abstract Time series forecasting models are important tools in supporting production planning and decision-making of the companies, such as agroindustries. One of the challenges facing Brazilian chicken agribusinesses, one of the most important in the world, is to correctly forecast the demand of its customers. In this context, the aim of this paper was to propose a parametric model for demand forecasting based on the data of dispatch of chicks of the broiler line in an agroindustry in the Midwest of Minas Gerais. Different methods were evaluated over the 59-week historical series in order to identify the behavior of the weekly expedition and to verify possible trends and seasonalities. Data of 56 weeks were evaluated and the candidate models were obtained from the analysis of the autocorrelations of the observations. Through the BIC and AIC evaluation criteria, the First Order Autoregressive model (AR) was the most appropriate one. To assess the predictive accuracy of the AR (1) model, a comparison was made between predicted and observed values in the last four weeks of the time-series. Through this analysis, a satisfactory performance was verified, since the observed values were within the 95% confidence interval constructed through the model. Keywords: Time-series. Demand forecasting. Broiler chicks. Chicken.

scholarly journals Forecasting peak energy demand for smart buildings

2020 ◽  
Author(s):  
Mona A. Alduailij ◽  
Ioan Petri ◽  
Omer Rana ◽  
Mai A. Alduailij ◽  
Abdulrahman S. Aldawood
Keyword(s):  
Neural Network ◽  
Time Series ◽  
Energy Consumption ◽  
Energy Demand ◽  
Arima Model ◽  
Demand Forecasting ◽  
User Preferences ◽  
Machine Learning Techniques ◽  
Peak Energy ◽  
The Impact

AbstractPredicting energy consumption in buildings plays an important part in the process of digital transformation of the built environment, and for understanding the potential for energy savings. This also contributes to reducing the impact of climate change, where buildings need to increase their adaptability and resilience while reducing energy consumption and maintain user comfort. The use of Internet of Things devices for monitoring and control of energy consumption in buildings can take into account user preferences, event monitoring and building optimization. Detecting peak energy demand from historical building data can enable users to manage their energy use more efficiently, while also enabling real-time response strategies (including control and actuation) to known or future scenarios. Several statistical, time series, and machine learning techniques are proposed in this work to predict electricity consumption for five different building types, by using peak demand forecasting to achieve energy efficiency. We have used several indigenous and exogenous variables with a view to test different energy forecasting scenarios. The suggested techniques are evaluated for creating predictive models, including linear Regression, dynamic regression, ARIMA time series, exponential smoothing time series, artificial neural network, and deep neural network. We conduct the analysis on an energy consumption dataset of five buildings from 2014 until 2019. Our results show that for a day ahead prediction, the ARIMA model outperforms the other approaches with an accuracy of 98.91% when executed over a 168 h (1 week) of uninterrupted data for five government buildings.

Sign in / Sign up

Export Citation Format

Share Document