1. Для TSTS schema

1.1 Load data:

ts <- read.csv("https://raw.githubusercontent.com/forvis/forvision_data/master/M3_quarterly_TSTS.csv")
head(ts)
##   series_id category   value timestamp
## 1        Q1    MICRO 3142.63   1984-Q1
## 2        Q1    MICRO 3190.75   1984-Q2
## 3        Q1    MICRO 3178.69   1984-Q3
## 4        Q1    MICRO 3170.94   1984-Q4
## 5        Q1    MICRO 3124.38   1985-Q1
## 6        Q1    MICRO 3170.00   1985-Q2

1.2 ValidateTSTS():

library(forvision)
validateTSTS(ts)
## [1] TRUE
  • Если нет какой-то колонки:
ts1 <- ts
ts1$timestamp <- NULL
validateTSTS(ts1)

Error in validateTSTS(ts1) : Check the column names of input data frame. The input data needed in the form of a data frame containing columns named ‘series_id’, ‘timestamp’, and ‘value’.

1.3 getSeriesSummary():

getSeriesSummary(ts)
##   Time series data summary
##   ========================
##   Number of time series:        756
##   Total number of observations: 37004
##   Timestamp range:              from 1946-Q1 to 1996-Q3

2. Для FTS schema:

2.1 Load data:

fs <- read.csv("https://raw.githubusercontent.com/forvis/forvision_data/master/M3_quarterly_FTS.csv")
head(fs)
##   series_id category method forecast horizon timestamp origin_timestamp
## 1        Q1    MICRO NAIVE2  5511.55       1   1993-Q1          1992-Q4
## 2        Q1    MICRO NAIVE2  5511.55       2   1993-Q2          1992-Q4
## 3        Q1    MICRO NAIVE2  5511.55       3   1993-Q3          1992-Q4
## 4        Q1    MICRO NAIVE2  5511.55       4   1993-Q4          1992-Q4
## 5        Q1    MICRO NAIVE2  5511.55       5   1994-Q1          1992-Q4
## 6        Q1    MICRO NAIVE2  5511.55       6   1994-Q2          1992-Q4

2.2 ValidateTSTS():

  • Колонки horizon не numeric:
library(forvision)
validateFTS(fs)

column horizon must be numeric

  • преобразование этой колонки в numeric:
fs$horizon <- as.numeric(fs$horizon)
validateFTS(fs)

2.3 getForecastSummary():

library(forvision)
#Convert column horizon to numeric
fs$horizon <- as.numeric(fs$horizon)
getForecastSummary(fs)
##   Forecast data summary
##   ========================
##   Number of time series:     756
##   Timestamp range:           from 1959-Q1 to 1996-Q3
##   Origin range:              from 1958-Q4 to 1994-Q3
##   Number of methods:         24
##   Total number of forecasts: 145152
##   Horizon range:             from 1 to 8