Probabilistic demand forecasting at scale

Böse, Joos-Hendrik; Flunkert, Valentín; Gasthaus, Jan; Januschowski, Tim; Lange, Dustin; Salinas, David; Schelter, Sebastian; Seeger, Matthias; Wang, Yuyang

doi:10.14778/3137765.3137775

Cited by 90 publications

(53 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Data validation in ML applications The challenges involved in building complex machine learning applications in the real-world have recently been highlighted by many researchers, e.g., with respect to managing the resulting models [26], software engineering [42,8], pipeline abstractions [2,43,46], and learnings from real-world systems [34,7]. A new focus is being put on data management questions with respect to large-scale machine learning.…”

Section: Related Workmentioning

confidence: 99%

“…Data is at the center of modern enterprises and institutions. Online retailers, for example, rely on data to support customers making buying decisions, to forecast demand [7], to schedule deliveries, and more generally, to guide every single business process and decision. Missing or incorrect information seriously compromises any decision process downstream, ultimately damaging the overall effectiveness and efficiency of the organization.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Automating large-scale data quality verification

et al. 2018

Self Cite

View full text Add to dashboard Cite

Modern companies and institutions rely on data to guide every single business process and decision. Missing or incorrect information seriously compromises any decision process downstream. Therefore, a crucial, but tedious task for everyone involved in data processing is to verify the quality of their data. We present a system for automating the verification of data quality at scale, which meets the requirements of production use cases. Our system provides a declarative API, which combines common quality constraints with userdefined validation code, and thereby enables 'unit tests' for data. We efficiently execute the resulting constraint validation workload by translating it to aggregation queries on Apache Spark. Our platform supports the incremental validation of data quality on growing datasets, and leverages machine learning, e.g., for enhancing constraint suggestions, for estimating the 'predictability' of a column, and for detecting anomalies in historic data quality time series. We discuss our design decisions, describe the resulting system architecture, and present an experimental evaluation on various datasets.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Automating large-scale data quality verification

et al. 2018

Self Cite

View full text Add to dashboard Cite

show abstract

“…A previsão da demanda é um dos principais desafios em qualquer negócio, sendo a chave para uma melhor gestão da empresa, com minimização de riscos devido a indisponibilidade de estoque ou ainda com custos menores devido a um inventário melhor planejado. Esta prática, constitui-se como uma previsão estimada por um produto ou lotes destes em uma faixa de tempo, considerando dados de entrada como registros históricos de vendas (Bose, Flunkert, Gasthaus & Januschowski, 2017).…”

Section: Previsão Da Demandaunclassified

Previsão da demanda como suporte à filosofia lean

Santos

2019

Exacta

View full text Add to dashboard Cite

O pensamento lean pode ser traduzido como um conjunto de princípios, ferramentas e técnicas que, combinadas, objetivam a melhoria, em termos de qualidade, custo, satisfação do cliente e outros. Diversas ferramentas e métodos podem auxiliar as organizações na busca por um ambiente fabril “enxuto”, desde por exemplo ferramentas da qualidade, tecnologia da informação e ainda, métodos matemático-estatísticos, como a previsão da demanda. Assim, o presente artigo que comporta-se como estudo de caso, teve como objetivo analisar a aplicação do modelo de Holt-Winters para Efeitos Sazonais Aditivos, numa empresa fabricante de polpa de frutas localizada em Pernambuco e comparar os ganhos obtidos sob a ótica da filosofia lean. Observou-se, entre outros ganhos, menor nível de estoques, tempo de espera reduzido e ainda menores custos. Ou seja, comprova-se que, quando aplicada a técnica de previsão da demanda, a gestão da empresa pode planejar melhor suas aquisições, determinar investimentos em ativos, contratação de mão de obra, ampliação da capacidade produtiva e ainda, estimar lucro da mesma.

show abstract

“…Conclusions and Lessons learned. We conclude the tutorial with a summary of the previous parts and share the lessons learned building the scalable forecasting system for retail and cloud resource within Amazon [1] and developing cloud-based deep learning forecasting algorithms such as DeepAR [6] in AWS SageMaker [13].…”

Section: 1mentioning

confidence: 99%

Forecasting big time series

et al. 2018

Self Cite

View full text Add to dashboard Cite

Time series forecasting is a key ingredient in the automation and optimization of business processes: in retail, deciding which products to order and where to store them depends on the forecasts of future demand in different regions; in cloud computing, the estimated future usage of services and infrastructure components guides capacity planning; and workforce scheduling in warehouses, call centers, factories requires forecasts of the future workload. Recent years have witnessed a paradigm shift in forecasting techniques and applications, from computer-assisted model-and assumptionbased to data-driven and fully-automated. This shift can be attributed to the availability of large, rich, and diverse time series data sources, posing unprecedented challenges to traditional time series forecasting methods. As such, how can we build statistical models to efficiently and effectively learn to forecast from large and diverse data sources? How can we leverage the statistical power of "similar" time series to improve forecasts in the case of limited observations? What are the implications for building forecasting systems that can handle large data volumes? The objective of this tutorial is to provide a concise and intuitive overview of the most important methods and tools available for solving large-scale forecasting problems. We review the state of the art in three related fields: (1) classical modeling of time series, (2) scalable tensor methods, and (3) deep learning for forecasting. Further, we share lessons learned from building scalable forecasting systems. While our focus is on providing an intuitive overview of the methods and practical issues, we also present technical details underlying these powerful tools.

show abstract

Probabilistic demand forecasting at scale

Cited by 90 publications

References 18 publications

Automating large-scale data quality verification

Automating large-scale data quality verification

Previsão da demanda como suporte à filosofia lean

Forecasting big time series

Contact Info

Product

Resources

About