2017
DOI: 10.7287/peerj.preprints.3190v1
|View full text |Cite
Preprint
|
Sign up to set email alerts
|

Forecasting at Scale

Abstract: Forecasting is a common data science task that helps organizations with capacity planning, goal setting, and anomaly detection. Despite its importance, there are serious challenges associated with producing reliable and high quality forecastsespecially when there are a variety of time series and analysts with expertise in time series modeling are relatively rare. To address these challenges, we describe a practical approach to forecasting "at scale" that combines configurable models with analyst-in-the-loop pe… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

2
410
0
7

Year Published

2018
2018
2022
2022

Publication Types

Select...
6
2
1

Relationship

0
9

Authors

Journals

citations
Cited by 290 publications
(419 citation statements)
references
References 6 publications
2
410
0
7
Order By: Relevance
“…In our seasonal decomposition step, we use TBATS as a deseasonalisation technique to extract the relevant seasonal components of a time series. We perform the seasonal extraction after fitting the TBATS model using the tbats function provided by the forecast package [33], [34] in R. 4) Prophet: Prophet is an automated forecasting framework developed by Taylor and Letham [37]. The main aim of this framework is to address the challenges involved in forecasting at Facebook, the employer of those authors at that time.…”
Section: Seasonal Decompositionmentioning
confidence: 99%
See 1 more Smart Citation
“…In our seasonal decomposition step, we use TBATS as a deseasonalisation technique to extract the relevant seasonal components of a time series. We perform the seasonal extraction after fitting the TBATS model using the tbats function provided by the forecast package [33], [34] in R. 4) Prophet: Prophet is an automated forecasting framework developed by Taylor and Letham [37]. The main aim of this framework is to address the challenges involved in forecasting at Facebook, the employer of those authors at that time.…”
Section: Seasonal Decompositionmentioning
confidence: 99%
“…We compare our developments against a collection of current state-of-the-art techniques in forecasting with multiple seasonal cycles. This includes Tbats [18], Prophet [37], and FFORMA [51]. We also use two variants of Dynamic-Harmonic-Regression [33] as the benchmarks.…”
Section: E Benchmarks and Lstm-msnet Variantsmentioning
confidence: 99%
“…(2) Prophet: Prophet [20] is a Bayesian nonlinear univariate generative model for time series forecasting which was proposed by Facebook in 2018. Like our method, Prophet is also a structural time series analysis method, which explicitly models the trend, seasonality, and event effects.…”
Section: A Experimental Setup and Baselinesmentioning
confidence: 99%
“…It is also possible that long-term seasonal, day of week, and time of day effects can influence the outcome of N-of-1 studies. Future versions of our model may incorporate parameters for these effects and fit them using methods akin to those of Prophet [15] or other Bayesian time-series models.…”
Section: Study Limitations and Future Workmentioning
confidence: 99%