rlgt

Rlgt is an R package for Bayesian Exponential Smoothing

https://github.com/cbergmeir/rlgt

Science Score: 36.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • CITATION.cff file
  • codemeta.json file
    Found codemeta.json file
  • .zenodo.json file
    Found .zenodo.json file
  • DOI references
  • Academic publication links
  • Committers with academic emails
    3 of 8 committers (37.5%) from academic institutions
  • Institutional organization owner
  • JOSS paper metadata
  • Scientific vocabulary similarity
    Low similarity (11.8%) to scientific vocabulary
Last synced: 10 months ago · JSON representation

Repository

Rlgt is an R package for Bayesian Exponential Smoothing

Basic Info
  • Host: GitHub
  • Owner: cbergmeir
  • Language: R
  • Default Branch: master
  • Size: 142 MB
Statistics
  • Stars: 22
  • Watchers: 4
  • Forks: 11
  • Open Issues: 2
  • Releases: 0
Created over 7 years ago · Last pushed about 1 year ago
Metadata Files
Readme

README.md

README

This package is based on code from Slawek Smyl to implement LGT, a local- and global trend exponential smoothing forecasting method using Rstan for model fitting.

Installation

Rlgt: R CMD build runs the "cleanup" script that runs roxygen and creates the source files. Then, R CMD INSTALL can be used to install the .tar.gz package.

RlgtLik: Implementation of LGT into code extracted from the "forecast" package.

TODOs

  1. We need a paragraph to say what is so good about the package/model? Main selling points? What can we do that nobody else can?

An implementation of various Bayesian Exponential Smoothing (ETS) models which have been found to outperform all of the original models in the M3 competition. These models include LGT (Local-Global Trend), SGT (Seasonal Global Trend), and their variations. The Bayesian model fitting is based on the RStan package.

Local-global trend, multiple seasonalities, external variables, high accuracy, prediction intervals, non-normal errors (fat tailed), heteroscedasticity, very flexible...

trend between multiplicative and additive (between linear and non-linear) seasonality is also generalized between multiplicative and additive.

  1. In the vignette, we need one/some of the examples from the demo and go through them step by step (?) Which examples are suitable?

We want 4 demo's in total: LGT SGT LGTREG SGTREG

For LGT and SGT, we can use data from the forecast package, or M3, or R base (AirPassengers).

We want one vignette with 4 sections. Do not use lynx in vignette, as seasonality is controversial. Or, if used, use it as 5th example.

LGTREG Runs LGT forecast with and without a regression component. lynx demo with the lynx dataset. SGTREG Runs SGT forecast with and without a regression component.

LGTM3 Runs through a subset of M3 yearly data, showing two methods of passing the non-seasonal series data. LGT&SGTM3 Using M3 series, it demos several possibilities or sub-versions of seasonal and non-seasonal models. LGT&SGT Implementation of LGT and SGT models on the R buit-in datasets. S2GTM4 It demos several ways of passing data to dual seasonality models, and several dual seasonality models, using M4 hourly data set. S2GTM4Hourlyparallel Parallel implementation of S2GT (and occasionally SGT) models on hourly time-series of M4-competition dataset. SGTM3 It uses quarterly subset of M3 data to demo ways of passing seasonal data. SGTM3parallel Parallel implementation of SGT model on monthly time-series data from M3-competition dataset. SGTM4Weekly_parallel Parallel implementation of SGT model with non-integer seasonality on weekly time-series data from M4-competition dataset.

  1. Checking the demos:

running the lynx demo, I get:

SAMPLING FOR MODEL 'SGT' NOW (CHAIN 1). Chain 1: Initialization between (-2, 2) failed after 100 attempts. [1] "Error in sampler$callsampler(argslist[[i]]) : Initialization failed." error occurred during calling the sampler; sampling not done

--> if I reduce the "seasonality" parameter it works

S2GT_M4 demo gives me the same error, doesn't run.

--> Seems to be a problem with newest Stan version

  1. Rlgt-package: I added some pointers to the demos but should be expanded with content discussed above.

  2. DESCRIPTION FILE: revise description with "main selling points"

  3. need to describe umcsent.example dataset in data.R See here how to document various datasets: https://github.com/robjhyndman/forecast/blob/master/R/data.R

Owner

  • Login: cbergmeir
  • Kind: user

GitHub Events

Total
  • Watch event: 2
  • Push event: 6
Last Year
  • Watch event: 2
  • Push event: 6

Committers

Last synced: about 3 years ago

All Time
  • Total Commits: 196
  • Total Committers: 8
  • Avg Commits per committer: 24.5
  • Development Distribution Score (DDS): 0.673
Past Year
  • Commits: 3
  • Committers: 2
  • Avg Commits per committer: 1.5
  • Development Distribution Score (DDS): 0.333
Top Committers
Name Email Commits
Erwin e****4@s****u 64
slawek s****s@h****k 64
Christoph Bergmeir c****r@g****m 24
To Wang Ng e****g@u****m 23
Christoph Bergmeir c****r@m****u 17
Alexander Dokumentov a****v@g****m 2
Andrew Johnson a****n@p****u 1
Chi Po Choi c****i@u****m 1
Committer Domains (Top 20 + Academic)

Issues and Pull Requests

Last synced: 11 months ago

All Time
  • Total issues: 3
  • Total pull requests: 4
  • Average time to close issues: about 19 hours
  • Average time to close pull requests: 3 days
  • Total issue authors: 3
  • Total pull request authors: 3
  • Average comments per issue: 3.0
  • Average comments per pull request: 0.75
  • Merged pull requests: 4
  • Bot issues: 0
  • Bot pull requests: 0
Past Year
  • Issues: 0
  • Pull requests: 0
  • Average time to close issues: N/A
  • Average time to close pull requests: N/A
  • Issue authors: 0
  • Pull request authors: 0
  • Average comments per issue: 0
  • Average comments per pull request: 0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
Top Authors
Issue Authors
  • AlbertoAlmuinha (1)
  • naodaihuang (1)
  • yonoklee (1)
Pull Request Authors
  • andrjohns (2)
  • pochoi (1)
  • jgabry (1)
Top Labels
Issue Labels
Pull Request Labels

Packages

  • Total packages: 2
  • Total downloads:
    • cran 708 last-month
  • Total docker downloads: 34
  • Total dependent packages: 0
    (may contain duplicates)
  • Total dependent repositories: 1
    (may contain duplicates)
  • Total versions: 9
  • Total maintainers: 1
cran.r-project.org: Rlgt

Bayesian Exponential Smoothing Models with Trend Modifications

  • Versions: 7
  • Dependent Packages: 0
  • Dependent Repositories: 1
  • Downloads: 708 Last month
  • Docker Downloads: 34
Rankings
Forks count: 6.3%
Stargazers count: 12.2%
Average: 19.9%
Docker downloads count: 20.1%
Dependent repos count: 24.0%
Downloads: 27.9%
Dependent packages count: 28.8%
Last synced: 11 months ago
conda-forge.org: r-rlgt
  • Versions: 2
  • Dependent Packages: 0
  • Dependent Repositories: 0
Rankings
Dependent repos count: 34.0%
Forks count: 40.9%
Average: 43.0%
Stargazers count: 45.8%
Dependent packages count: 51.2%
Last synced: 10 months ago

Dependencies

Rlgt/DESCRIPTION cran
  • R >= 3.4.0 depends
  • Rcpp >= 0.12.0 depends
  • forecast * depends
  • methods * depends
  • rstantools * depends
  • truncnorm * depends
  • rstan >= 2.18.1 imports
  • sn * imports
  • knitr * suggests
  • rmarkdown * suggests
RlgtLik/DESCRIPTION cran
  • R >= 3.0.2 depends
  • forecast * depends
  • Rcpp >= 0.11.0 imports
  • testthat * suggests