DeepAR: Probabilistic forecasting with autoregressive recurrent networks

Probabilistic forecasting, i.e., estimating a time series’ future probability distribution given its past, is a key enabler for optimizing business processes. In retail businesses, for example, probabilistic demand forecasts are crucial for having the right inventory available at the right time and in the right place. This paper proposes DeepAR, a methodology for producing accurate probabilistic forecasts, based on training an autoregressive recurrent neural network model on a large number of related time series. We demonstrate how the application of deep learning techniques to forecasting can overcome many of the challenges that are faced by widely-used classical approaches to the problem. By means of extensive empirical evaluations on several real-world forecasting datasets, we show that our methodology produces more accurate forecasts than other state-of-the-art methods, while requiring minimal manual work.     

CPFR中的联合预测研究

文章研究了联合计划、预测和补货（CPFR）中的联合预测流程,并建立了相关的预测模型。在建模的过程中,使用了状态空间方程来描述实际市场需求和观测到的市场需求（销售量）,并通过卡尔曼滤波来预测零售商下期的销售量．结合零售商库存策略,预测出零售商下期的订单量。     

波动性产品需求预测的神经网络模型

随着市场竞争的日益激烈,消费者的心里越来越复杂,这样导致了产品的需求的波动性大大增加。这种强波动性的产品需求序列中除了随机性外还存在混沌性,根据混沌理论可知,混沌的短期预测是可行的。为了有效的对这些混沌性进行预测,选择了神经网络作为预测模型,因为神经网络对非线性具有较好逼近能力。在网络结构选择中考虑了混沌序列的嵌入维数,并在隐层中加入了径向基以更好的拟合数据。在针对目前很多企业具有数据库和数据仓库的背景,给出了基于数据挖掘的具体预测方法,并通过实例演示了预测的有效性。     

Stability in the inefficient use of forecasting systems: A case study in a supply chain company

Computer-based demand forecasting systems have been widely adopted in supply chain companies, but little research has studied how these systems are actually used in the forecasting process. We report the findings of a case study of demand forecasting in a pharmaceutical company over a 15-year period. At the start of the study, managers believed that they were making extensive use of their forecasting system that was marketed based on the accuracy of its advanced statistical methods. Yet most forecasts were obtained using the system’s facility for judgmentally overriding the automatic statistical forecasts. Carrying out the judgmental interventions involved considerable management effort as part of a sales & operations planning (S&OP) process, yet these often only served to reduce forecast accuracy. This study uses observations of the forecasting process, interviews with participants and data on the accuracy of forecasts to investigate why the managers continued to use non-normative forecasting practices for many years despite the potential economic benefits that could be achieved through change. The reasons for the longevity of these practices are examined both from the perspective of the individual forecaster and the organization as a whole.     

具有任意投资水平的供应链协同预测博弈研究

运用纳什均衡和贝叶斯更新模型,得到了供应链联合预测均衡的存在条件。模型中,供应商和零售商均需决定在预测技术上的投资水平,双方的需求预测将会被汇总成一个统一的预测。结果表明,双方预测能力越接近中等水平,越容易实现联合预测。预测能力偏离中等水平越远,越容易出现搭便车行为,即至少有一方不进行预测。     

Anticipating special events in Emergency Department forecasting

Accurate daily forecast of Emergency Department (ED) attendance helps roster planners in allocating available resources more effectively and potentially influences staffing. Since special events affect human behaviours, they may increase or decrease the demand for ED services. Therefore, it is crucial to model their impact and use them to forecast future attendance to improve roster planning and avoid reactive strategies. In this paper, we propose, for the first time, a forecasting model to generate both point and probabilistic daily forecast of ED attendance. We model the impact of special events on ED attendance by considering real-life ED data. We benchmark the accuracy of our model against three time-series techniques and a regression model that does not consider special events. We show that the proposed model outperforms its benchmarks across all horizons for both point and probabilistic forecasts. Results also show that our model is more robust with an increasing forecasting horizon. Moreover, we provide evidence on how different types of special events may increase or decrease ED attendance. Our model can easily be adapted for use not only by EDs but also by other health services. It could also be generalised to include more types of special events.     

信息分享对原始设备制造商与其合同制造商竞争的影响

研究了原始设备制造商的预测信息分享对一个原始设备制造商和一个与其同时有合作和竞争的合同制造商组成的供应链系统的影响,建立制造商间信息分享的模型,该模型包括一个原始设备制造商和一个合同制造商。研究发现,原始设备制造商关于市场潜在需求预测信息的分享对其预期利润是不利的,同时需求信息预测的精度对原始设备制造商信息分享的决策也有影响,原始设备制造商没有动机与其供应链成员进行信息分享,但信息分享使得供应链整体利润增加。最后,建立一个信息分享补偿机制分享供应链利润的增加量,以期通过信息分享补偿机制促使原始设备制造商有动机进行信息分享,从而实现其与合同制造商的“双赢”。     

Forecasting sales in the supply chain: Consumer analytics in the big data era

Forecasts have traditionally served as the basis for planning and executing supply chain activities. Forecasts drive supply chain decisions, and they have become critically important due to increasing customer expectations, shortening lead times, and the need to manage scarce resources. Over the last ten years, advances in technology and data collection systems have resulted in the generation of huge volumes of data on a wide variety of topics and at great speed. This paper reviews the impact that this explosion of data is having on product forecasting and how it is improving it. While much of this review will focus on time series data, we will also explore how such data can be used to obtain insights into consumer behavior, and the impact of such data on organizational forecasting.     

Demand forecasting and order planning for humanitarian logistics: An empirical assessment

Humanitarian aid organizations are most known for their short-term emergency relief. While getting aid items to those in need can be challenging, long-term projects provide an opportunity for demand planning supported by forecasting methods. Based on standardized consumption data of the Operational Center Amsterdam of Médecins Sans Frontières (MSF-OCA) regarding nineteen longer-term aid projects and over 2000 medical items consumed in 2013, we describe and analyze the forecasting and order planning process. We find that several internal and external factors influence forecast and order planning performance, be it indirectly through demand volatility and safety markup. Moreover, we identify opportunities for further improvement for MSF-OCA, and for humanitarian logistics organizations in general.     

基于小波变换和支持向量的国际原油价格预测

本文将小波分析与支持向量回归结合应用于国际原油价格预测,通过小波多尺度分析方法将油价时间序列分解为长期趋势和随机扰动项,然后采用支持向量回归对分解后的油价长期趋势进行预测。油价长期趋势的预测采用多因素预测方法,主要考虑市场供需基本面、库存、经济、投机等因素对石油价格走势的影响,建立多输入单输出的支持向量回归模型。实证研究表明,支持向量回归模型具有较高的预测性能,对原油价格长期趋势预测中,该方法比回归方法的预测精度高。     

Comparison of intraday probabilistic forecasting of solar irradiance using only endogenous data

Accurate solar forecasts are necessary to improve the integration of solar renewables into the energy grid. In recent years, numerous methods have been developed for predicting the solar irradiance or the output of solar renewables. By definition, a forecast is uncertain. Thus, the models developed predict the mean and the associated uncertainty. Comparisons are therefore necessary and useful for assessing the skill and accuracy of these new methods in the field of solar energy.The aim of this paper is to present a comparison of various models that provide probabilistic forecasts of the solar irradiance within a very strict framework. Indeed, we consider focusing on intraday forecasts, with lead times ranging from 1 to 6 h. The models selected use only endogenous inputs for generating the forecasts. In other words, the only inputs of the models are the past solar irradiance data. In this context, the most common way of generating the forecasts is to combine point forecasting methods with probabilistic approaches in order to provide prediction intervals for the solar irradiance forecasts. For this task, we selected from the literature three point forecasting models (recursive autoregressive and moving average (ARMA), coupled autoregressive and dynamical system (CARDS), and neural network (NN)), and seven methods for assessing the distribution of their error (linear model in quantile regression (LMQR), weighted quantile regression (WQR), quantile regression neural network (QRNN), recursive generalized autoregressive conditional heteroskedasticity (GARCHrls), sieve bootstrap (SB), quantile regression forest (QRF), and gradient boosting decision trees (GBDT)), leading to a comparison of 20 combinations of models.None of the model combinations clearly outperform the others; nevertheless, some trends emerge from the comparison. First, the use of the clear sky index ensures the accuracy of the forecasts. This derived parameter permits time series to be deseasonalized with missing data, and is also a good explanatory variable of the distribution of the forecasting errors. Second, regardless of the point forecasting method used, linear models in quantile regression, weighted quantile regression and gradient boosting decision trees are able to forecast the prediction intervals accurately.     

Big data from dynamic pricing: A smart approach to tourism demand forecasting

Suppliers of tourist services continuously generate big data on ask prices. We suggest using this information, in the form of a price index, to forecast the occupation rates for virtually any time-space frame, provided that there are a sufficient number of decision makers “sharing” their pricing strategies on the web. Our approach guarantees great transparency and replicability, as big data from OTAs do not depend on search interfaces and can facilitate intelligent interactions between the territory and its inhabitants, thus providing a starting point for a smart decision-making process. We show that it is possible to obtain a noticeable increase in the forecasting performance by including the proposed leading indicator (price index) into the set of explanatory variables, even with very simple model specifications. Our findings offer a new research direction in the field of tourism demand forecasting leveraging on big data from the supply side.     

Improving demand forecasting in the air cargo handling industry: a case study

Air transportation plays a crucial role in the agile and dynamic environment of contemporary supply chains. This industry is characterised by high air cargo demand uncertainty, making forecasting extremely challenging. An in-depth case study has been undertaken in order to explore and untangle the factors influencing demand forecasting and consequently to improve the operational performance of an air cargo handling company. It has been identified that in practice, the demand forecasting process does not provide the necessary level of accuracy, to effectively cope with the high demand uncertainty. This has a negative impact on a whole range of air cargo operations, but especially on the management of the workforce, which is the most expensive resource in the air cargo handling industry. Besides forecast inaccuracy, a range of additional hidden factors that affect operations management have been identified. A number of recommendations have been made to improve demand forecasting and workforce management.     

A fast and scalable ensemble of global models with long memory and data partitioning for the M5 forecasting competition

This work presents key insights on the model development strategies used in our cross-learning-based retail demand forecast framework. The proposed framework outperforms state-of-the-art univariate models in the time series forecasting literature. It has achieved 17th position in the accuracy track of the M5 forecasting competition, which is among the top 1% of solutions.     

Forecasting the UK economy with a medium-scale Bayesian VAR

We estimate a Bayesian VAR (BVAR) for the UK economy and assess its performance in forecasting GDP growth and CPI inflation in real time relative to forecasts from COMPASS, the Bank of England’s DSGE model, and other benchmarks. We find that the BVAR outperformed COMPASS when forecasting both GDP and its expenditure components. In contrast, their performances when forecasting CPI were similar. We also find that the BVAR density forecasts outperformed those of COMPASS, despite under-predicting inflation at most forecast horizons. Both models over-predicted GDP growth at all forecast horizons, but the issue was less pronounced in the BVAR. The BVAR’s point and density forecast performances are also comparable to those of a Bank of England in-house statistical suite for both GDP and CPI inflation, as well as to the official Inflation Report projections. Our results are broadly consistent with the findings of similar studies for other advanced economies.     

Design and development of inventory knowledge discovery system

Inventory management (IM) performance is affected by the forecasting accuracy of both demand and supply. In this paper, an inventory knowledge discovery system (IKDS) is designed and developed to forecast and acquire knowledge among variables for demand forecasting. In IKDS, the TREes PArroting Networks (TREPAN) algorithm is used to extract knowledge from trained networks in the form of decision trees which can be used to understand previously unknown relationships between the input variables so as to improve the forecasting performance for IM. The experimental results show that the forecasting accuracy using TREPAN is superior to traditional methods like moving average and autoregressive integrated moving average. In addition, the knowledge extracted from IKDS is represented in a comprehensible way and can be used to facilitate human decision-making.     

Integrated manufacturing/distribution planning decisions with multiple imprecise goals in an uncertain environment

This work presents a possibilistic linear programming (PLP) method for solving the integrated manufacturing/distribution planning decision (MDPD) problems with multiple imprecise goals in supply chains under an uncertain environment. The imprecise PLP model designed here aims to simultaneously minimize total net costs and total delivery time with reference to available supply, capacities, labor levels, quota flexibility and cost budget constraints at each source, as well as forecast demand and warehouse space at each destination. The proposed method achieves greater computational efficiency by employing the simplified triangular distribution to represent imprecise numbers. An industrial case is used to demonstrate the feasibility of applying the proposed method to a real MDPD problem. Overall, the proposed PLP method provides a practical means of solving the multi-objective MDPD problems in an uncertain environment, and can effectively improve manufacturer/ distributor relationships in a supply chain.     

Daily retail demand forecasting using machine learning with emphasis on calendric special days

Demand forecasting is an important task for retailers as it is required for various operational decisions. One key challenge is to forecast demand on special days that are subject to vastly different demand patterns than on regular days. We present the case of a bakery chain with an emphasis on special calendar days, for which we address the problem of forecasting the daily demand for different product categories at the store level. Such forecasts are an input for production and ordering decisions. We treat the forecasting problem as a supervised machine learning task and provide an evaluation of different methods, including artificial neural networks and gradient-boosted decision trees. In particular, we outline and discuss the possibility of formulating a classification instead of a regression problem. An empirical comparison with established approaches reveals the superiority of machine learning methods, while classification-based approaches outperform regression-based approaches. We also found that machine learning methods not only provide more accurate forecasts but are also more suitable for applications in a large-scale demand forecasting scenario that often occurs in the retail industry.     

基于BP神经网络与GM(1,1)组合的北京市物流需求预测模型

通过分析影响北京市物流需求的相关因素,构建北京市物流需求预测影响因素指标体系。运用BP神经网络和GM(1,1)方法,建立北京市物流需求组合预测模型,选取近20年的统计数据对未来五年的物流需求进行预测,得出物流需求总量及变化规律,并以此提出推进北京市物流业发展的有效途径,为物流系统规划提供合理依据及有效发展途径。     

Empirical prediction intervals revisited

Empirical prediction intervals are constructed based on the distribution of previous out-of-sample forecast errors. Given historical data, a sample of such forecast errors is generated by successively applying a chosen point forecasting model to a sequence of fixed windows of past observations and recording the associated deviations of the model predictions from the actual observations out-of-sample. The suitable quantiles of the distribution of these forecast errors are then used along with the point forecast made by the selected model to construct an empirical prediction interval. This paper re-examines the properties of the empirical prediction interval. Specifically, we provide conditions for its asymptotic validity, evaluate its small sample performance and discuss its limitations.