Forecasting at Scale

Taylor, Sean J.; Letham, Benjamin

doi:10.1080/00031305.2017.1380080

Cited by 1,319 publications

(518 citation statements)

References 14 publications

Supporting

Mentioning

379

Contrasting

Unclassified

Order By: Relevance

“…(ii) Time series decomposition: distinction of the seasonal (periodic) components of the time series was performed with the Prophet model (Taylor and Letham ). If y ( t ) is the observed time series, Prophet uses an additive model, where linear growth trends ( g ( t )) are fitted with periodic seasonalities ( s ( t )), plus a normal random error term ( ϵ ( t )):

y (t) = g (t) + s (t) + ϵ (t)

…”

Section: Methodsmentioning

confidence: 99%

“…The Prophet developers set default values for the parameters that are appropriate for most forecasting problems (Taylor and Letham 2018a): τ = 0.05, σ 2 = 10, 10 Fourier terms for the yearly periodic component, three Fourier terms for weekly periodic component and four Fourier terms for daily periodic component. These values proved to be suitable for our data.…”

Section: Methodsmentioning

confidence: 99%

“…Time series decomposition is another approach used to characterize hydrological patterns (Machiwal and Jha 2006;Von Asmuth et al 2008;Peterson and Western 2014;Wang et al 2015;Chiaudani et al 2017;Haaf and Barthel 2018). Concerning this method, there is a new flexible additive model developed by Facebook (Prophet) that considers non-periodic changes in trends as well as customizable seasonal periodic components in a Bayesian framework as easily interpretable parameters (Taylor and Letham 2018a). This model's applicability to hydrometeorological data has already been highlighted by some authors such as Papacharalampous et al (2018).…”

Section: Ngwaorgmentioning

confidence: 99%

See 2 more Smart Citations

Unraveling the Hydrological Behavior of a Coastal Pond in Doñana National Park (Southwest Spain)

Fernández-Ayuso¹,

Aguilera²,

Guardiola‐Albert³

et al. 2019

Groundwater

View full text Add to dashboard Cite

Time series analysis methods have been used to detect behavioral patterns in a set of nine time series. These series contained information in a 3‐h time step about meteorological, hydrological and tidal data of a sand dune pond area located in Doñana National Park in the southwest of Spain. The methods used, such as wavelet analysis and additive seasonal decomposition, had never been applied before in the types of ecosystems studied. These approaches have improved the current knowledge of the conceptual model of the Santa Olalla pond system, the only system with a permanent hydroperiod located in this protected area. In addition, complex surface water‐groundwater interactions, not visible through descriptive methods, have been distinguished to have a strong seasonal component. Finally, we evaluated the effect of pumping activity in a nearby coastal resort on the water supply of the Santa Olalla pond system. Although direct damage to this sand dune pond has not yet been identified, special attention must be paid in order to maintain groundwater inputs that are integral to maintaining its current status.

show abstract

y (t) = g (t) + s (t) + ϵ (t)

…”

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Ngwaorgmentioning

confidence: 99%

See 1 more Smart Citation

Unraveling the Hydrological Behavior of a Coastal Pond in Doñana National Park (Southwest Spain)

Fernández-Ayuso¹,

Aguilera²,

Guardiola‐Albert³

et al. 2019

Groundwater

View full text Add to dashboard Cite

show abstract

“…Prophet is robust to missing data and shifts in the trend, and typically handles outliers well. According to Taylor and Letham [23] research, Prophet is used in many applications across Facebook for producing reliable forecasts and performs better than any other approach in the majority of cases.…”

Section: Prophet -A Tool For Time Series Forecasting At Scalementioning

confidence: 99%

Application of Facebook's Prophet Algorithm for Successful Sales Forecasting Based on Real-world Data

Žunić¹,

Korjenic²,

Hodžić³

et al. 2020

IJCSIT

View full text Add to dashboard Cite

This paper presents a framework capable of accurately forecasting future sales in the retail industry and classifying the product portfolio according to the expected level of forecasting reliability. The proposed framework, that would be of great use for any company operating in the retail industry, is based on Facebook's Prophet algorithm and backtesting strategy. Real-world sales forecasting benchmark data obtained experimentally in a production environment in one of the biggest retail companies in Bosnia and Herzegovina is used to evaluate the framework and demonstrate its capabilities in a real-world use case scenario.

show abstract

“…En esa línea, podemos observar que, sobre todo en países de Europa central y en Estados Unidos, se ha comenzado a testear científicamente la capacidad predictiva de las RR.SS., especialmente en contextos electorales y también en estudios de mercado (Taylor & Lethman, 2017). Ya no solamente se trata del uso de la IA para la comunicación política, se avanza un paso y se explora la red para predecir resultados electorales (Schoen et al, 2013).…”

Section: Leer Las Redes Leer La Sociedad: Uso Y Estudio De Las Rrssunclassified

Redes sociales, inteligencia computacional y predicción electoral: el caso de las primarias presidenciales de Chile 2017

Santander

Gómez

González

et al. 2018

Cuadernos de información

View full text Add to dashboard Cite

show abstract

Forecasting at Scale

Cited by 1,319 publications

References 14 publications

Unraveling the Hydrological Behavior of a Coastal Pond in Doñana National Park (Southwest Spain)

Unraveling the Hydrological Behavior of a Coastal Pond in Doñana National Park (Southwest Spain)

Application of Facebook's Prophet Algorithm for Successful Sales Forecasting Based on Real-world Data

Redes sociales, inteligencia computacional y predicción electoral: el caso de las primarias presidenciales de Chile 2017

Contact Info

Product

Resources

About