| Type: | Package |
| Title: | Automatic Model Selection and Prediction for Univariate Time Series |
| Version: | 0.9.11 |
| Author: | Vivien Roussez |
| Maintainer: | Vivien Roussez <vivien.roussez@gmail.com> |
| Description: | Offers a set of functions to easily make predictions for univariate time series. 'autoTS' is a wrapper of existing functions of the 'forecast' and 'prophet' packages, harmonising their outputs in tidy dataframes and using default values for each. The core function getBestModel() allows the user to effortlessly benchmark seven algorithms along with a bagged estimator to identify which one performs the best for a given time series. |
| License: | GPL-3 |
| Encoding: | UTF-8 |
| LazyData: | true |
| Imports: | rlang, prophet, dplyr, magrittr, lubridate, tidyr, forecast, ggplot2, RcppRoll,shiny, shinycssloaders, plotly |
| BugReports: | https://github.com/vivienroussez/autots/issues |
| URL: | https://github.com/vivienroussez/autoTS |
| Suggests: | knitr,rmarkdown,stringr |
| VignetteBuilder: | knitr |
| RoxygenNote: | 7.1.0 |
| NeedsCompilation: | no |
| Packaged: | 2020-06-05 11:47:28 UTC; vivienroussez |
| Repository: | CRAN |
| Date/Publication: | 2020-06-05 12:20:06 UTC |
Creates additional dates and values when NA where removed and the TS is not complete
Description
Creates additional dates and values when NA where removed and the TS is not complete
Usage
complete.ts(dates, values, freq, complete = 0)
Arguments
dates |
A vector of dates that can be parsed by lubridate |
values |
A vector of same size as |
freq |
A chacracter string that indicates the frequency of the time series ("week", "month", "quarter", "day"). |
complete |
A numerical value (or NA) to fill the missing data points |
Value
A dataframe with 2 columns : date and val, with additional rows
Examples
library(lubridate)
library(dplyr)
dates <- seq(as_date("2000-01-01"),as_date("2010-12-31"),"month")
values <- rnorm(length(dates))
complete.ts(dates,values,"month",complete = 0)
Determine best algorithm
Description
Implement selected algorithms, train them without the last n observed data points (or n_test number of points), and compares the results to reality to determine the best algorithm
Usage
getBestModel(
dates,
values,
freq,
complete = 0,
n_test = NA,
graph = TRUE,
algos = list("my.prophet", "my.ets", "my.sarima", "my.tbats", "my.bats", "my.stlm",
"my.shortterm"),
bagged = "auto",
metric.error = my.rmse
)
Arguments
dates |
A vector of dates that can be parsed by lubridate |
values |
A vector of same size as |
freq |
A chacracter string that indicates the frequency of the time series ("week", "month", "quarter", "day"). |
complete |
A numerical value (or NA) to fill the missing data points |
n_test |
number of data points to keep aside for the test (default : one year) |
graph |
A boolean, if TRUE, comparison of algorithms is plotted |
algos |
A list containing the algorithms (strings, with prefix "my.") to be tested |
bagged |
A string. "auto" will use all available algoriths, skipping algos parameter. Else, specified algos of the 'algo' parameter will be used |
metric.error |
a function to compute the error the each models. available functions : my.rmse and my.mae |
Value
A list contraining a character string with the name of the best method, a gg object with the comparison between algorithms and a dataframe with predictions of all tried algorithms, a dtaframe containing the errors of each algorithms, the preparedTS object and the list of algorithms tested
Examples
library(autoTS)
dates <- seq(lubridate::as_date("2005-01-01"),lubridate::as_date("2010-12-31"),"quarter")
values <- 10+ 1:length(dates)/10 + rnorm(length(dates),mean = 0,sd = 10)
which.model <- getBestModel(dates,values,freq = "quarter",n_test = 4)
### Custom set of algorithm (including for bagged estimator)
which.model <- getBestModel(dates,values,freq = "quarter",n_test = 4,
algos = list("my.prophet","my.ets"),bagged = "custom")
### Use MAE instead of RMSE
which.model <- getBestModel(dates,values,freq = "quarter",n_test = 3,
algos = list("my.prophet","my.ets"),
bagged = "custom",metric.error = my.mae)
Determines the decimal frequency of a time series from a character string
Description
Determines the decimal frequency of a time series from a character string
Usage
getFrequency(freq.alpha)
Arguments
freq.alpha |
A character string that indicates the frequency of the time series ("week", "month", "quarter", "day"). |
Value
The decimal version of the frequency (useful for the forecast package functions).
Examples
getFrequency("week")
Fit BATS algorithm and make the prediction
Description
Fit BATS algorithm and make the prediction
Usage
my.bats(prepedTS, n_pred)
Arguments
prepedTS |
A list created by the |
n_pred |
Int number of periods to forecast forward (eg n_pred = 12 will lead to one year of prediction for monthly time series) |
Value
A dataframe with 4 columns : date, average prediction, upper and lower 95
Examples
library(lubridate)
library(dplyr)
dates <- seq(as_date("2000-01-01"),as_date("2010-12-31"),"quarter")
values <- rnorm(length(dates))
my.ts <- prepare.ts(dates,values,"quarter",complete = 0)
my.bats(my.ts,n_pred=4)
Fit ETS algorithm and make the prediction
Description
Fit ETS algorithm and make the prediction
Usage
my.ets(prepedTS, n_pred)
Arguments
prepedTS |
A list created by the |
n_pred |
Int number of periods to forecast forward (eg n_pred = 12 will lead to one year of prediction for monthly time series) |
Value
A dataframe with 4 columns : date, average prediction, upper and lower 95
Examples
library(lubridate)
library(dplyr)
dates <- seq(as_date("2000-01-01"),as_date("2010-12-31"),"quarter")
values <- rnorm(length(dates))
my.ts <- prepare.ts(dates,values,"quarter",complete = 0)
my.ets(my.ts,n_pred=4)
Custom (internal) function for MAE
Description
Custom (internal) function for MAE
Usage
my.mae(true, predicted)
Arguments
true |
num vector of actual values |
predicted |
num vector of predicted values |
Value
Num value with MAE
Make predictions with selected algorithms
Description
Fit selected algorithms, make the predictions and combine the results along with observed data in one final dataframe.
Usage
my.predictions(
bestmod = NULL,
prepedTS = NULL,
algos = list("my.prophet", "my.ets", "my.sarima", "my.tbats", "my.bats", "my.stlm",
"my.shortterm"),
n_pred = NA
)
Arguments
bestmod |
A list produced by the |
prepedTS |
A list created by the |
algos |
A list containing the algorithms to be implemented. If |
n_pred |
Int number of periods to forecast forward (eg n_pred = 12 will lead to one year of prediction for monthly time series) |
Value
A dataframe containing : date, actual observed values, one column per used algorithm, and a column indicating the type of measure (mean prediction, upper or lower bound of CI)
Examples
library(lubridate)
library(dplyr)
dates <- seq(lubridate::as_date("2000-01-01"),lubridate::as_date("2010-12-31"),"quarter")
values <- 10+ 1:length(dates)/10 + rnorm(length(dates),mean = 0,sd = 10)
### Stand alone usage
prepare.ts(dates,values,"quarter") %>%
my.predictions(prepedTS = .,algos = list("my.prophet","my.ets"))
### Standard input with bestmodel
getBestModel(dates,values,freq = "quarter",n_test = 6) %>%
my.predictions()
Fit prophet algorithm and make the prediction
Description
Fit prophet algorithm and make the prediction
Usage
my.prophet(prepedTS, n_pred)
Arguments
prepedTS |
A list created by the |
n_pred |
Int number of periods to forecast forward (eg n_pred = 12 will lead to one year of prediction for monthly time series) |
Value
A dataframe for "next year" with 4 columns : date, average prediction, upper and lower 95
Examples
library(lubridate)
library(dplyr)
dates <- seq(as_date("2000-01-01"),as_date("2010-12-31"),"quarter")
values <- rnorm(length(dates))
my.ts <- prepare.ts(dates,values,"quarter",complete = 0)
my.prophet(my.ts,n_pred=4)
Custom (internal) function for RMSE
Description
Custom (internal) function for RMSE
Usage
my.rmse(true, predicted)
Arguments
true |
num vector of actual values |
predicted |
num vector of predicted values |
Value
Num value with RMSE
Fit SARIMA algorithm and make the prediction
Description
Fit SARIMA algorithm and make the prediction
Usage
my.sarima(prepedTS, n_pred)
Arguments
prepedTS |
A list created by the |
n_pred |
Int number of periods to forecast forward (eg n_pred = 12 will lead to one year of prediction for monthly time series) |
Value
A dataframe with 4 columns : date, average prediction, upper and lower 95
Examples
library(lubridate)
library(dplyr)
dates <- seq(as_date("2000-01-01"),as_date("2010-12-31"),"quarter")
values <- rnorm(length(dates))
my.ts <- prepare.ts(dates,values,"quarter",complete = 0)
my.sarima(my.ts,n_pred=4)
Fit short term algorithm and make the prediction
Description
Fit short term algorithm and make the prediction
Usage
my.shortterm(prepedTS, n_pred, smooth_window = 2)
Arguments
prepedTS |
A list created by the |
n_pred |
Int number of periods to forecast forward (eg n_pred = 12 will lead to one year of prediction for monthly time series). Note that this algorithm cannot predict further than one year |
smooth_window |
Int specifying the number of periods to consider for computing the evolution rate that will be applied for the forecast |
Details
this algorithm uses data of the last year and makes the prediction taking into account the seasonality and the evolution of the previous periods' evolution
Value
A dataframe with 4 columns : date, average prediction, upper and lower 95
Examples
library(lubridate)
library(dplyr)
dates <- seq(as_date("2000-01-01"),as_date("2010-12-31"),"quarter")
values <- rnorm(length(dates))
my.ts <- prepare.ts(dates,values,"quarter",complete = 0)
my.shortterm(my.ts,n_pred=4)
Fit STLM algorithm and make the prediction
Description
Fit STLM algorithm and make the prediction
Usage
my.stlm(prepedTS, n_pred)
Arguments
prepedTS |
A list created by the |
n_pred |
Int number of periods to forecast forward (eg n_pred = 12 will lead to one year of prediction for monthly time series) |
Value
A dataframe with 4 columns : date, average prediction, upper and lower 95
Examples
library(lubridate)
library(dplyr)
dates <- seq(as_date("2000-01-01"),as_date("2010-12-31"),"quarter")
values <- rnorm(length(dates))
my.ts <- prepare.ts(dates,values,"quarter",complete = 0)
my.stlm(my.ts,n_pred=4)
Fit TBATS algorithm and make the prediction
Description
Fit TBATS algorithm and make the prediction
Usage
my.tbats(prepedTS, n_pred)
Arguments
prepedTS |
A list created by the |
n_pred |
Int number of periods to forecast forward (eg n_pred = 12 will lead to one year of prediction for monthly time series) |
Value
A dataframe with 4 columns : date, average prediction, upper and lower 95
Examples
library(lubridate)
library(dplyr)
dates <- seq(as_date("2000-01-01"),as_date("2010-12-31"),"quarter")
values <- rnorm(length(dates))
my.ts <- prepare.ts(dates,values,"quarter",complete = 0)
my.tbats(my.ts,n_pred=4)
Format 2 vectors in a proper object usable by all algorithms
Description
Format 2 vectors in a proper object usable by all algorithms
Usage
prepare.ts(dates, values, freq, complete = 0)
Arguments
dates |
A vector of dates that can be parsed by lubridate |
values |
A vector of same size as |
freq |
A chacracter string that indicates the frequency of the time series ("week", "month", "quarter", "day"). |
complete |
A numerical value (or NA) to fill the missing data points |
Details
Creates a list with the time series in a dataframe and a ts object, and the frequency stored in decimal and litteral values. The result is meant to be put in the prophet or forecast functions
Value
A list containing : a dataframe, a ts vector for the time series, and 2 scalars for its frequency
Examples
library(lubridate)
library(dplyr)
library(ggplot2)
dates <- seq(lubridate::as_date("2000-01-01"),lubridate::as_date("2010-12-31"),"quarter")
values <- rnorm(length(dates))
my.ts <- prepare.ts(dates,values,"month",complete = 0)
plot(my.ts$obj.ts)
ggplot(my.ts$obj.df,aes(dates,val)) + geom_line()
Demo graphical user interface
Description
A shiny application that allows the user to load a properly formated CSV file, benchmark the algorithms, make a prediction and download the results. Requires additional packages shiny, shinycssloaders, tidyr and plotly to be installed
Usage
runUserInterface()
Examples
autoTS::runUserInterface()