Descomposición de Series Temporales
library(forecast)
library(readxl)
DatosIVAE <- read_excel("C:/Users/MINEDUCYT/Downloads/DatosIVAE.xlsx",
col_type = c("skip", "numeric"),
skip = 5)
Serie.IMAE.ts <- ts(data = DatosIVAE,
start = c(2018, 1),
frequency = 12)
Serie.IMAE.ts %>%
autoplot(main = "IMAE El Salvador 2018-2024[Marzo]",
xlab ="Años/Meses", ylab = "Indice")
Modelo aditivo.
- Serie temporal. \[Y_t = T_t+C_t + S_t + I_t\]
Componente Tendencial.
library(forecast)
Tta <- ma(Serie.IMAE.ts,12, centre = TRUE)
autoplot(Serie.IMAE.ts, main = "IMAE El Salvador 2018 - 2024[Marzo]",
xlab = "Años/Meses" , ylab = "Indice") +
autolayer(Tta, serie= "Tt")
Componente Estacional.
\[S_t = Y_t-T_t\]
library(magrittr)
Yt <- Serie.IMAE.ts
SIa <- Yt - Tta
Sta <- tapply(SIa, cycle(SIa), mean, na.rm = TRUE)
Sta <- Sta - sum(Sta)/12
Sta <- rep(Sta, len = length(Yt))%>%
ts(start = c(2018,1), frequency = 12)
autoplot(Sta, main= "Componente Estacional",
xlab ="Años/Meses", ylab ="Factores Estacionales")
Componente Irregular.
\[I_t= Y_t-T_t-S_t\]
library(forecast)
Ita <- Yt-Tta- Sta
autoplot(Ita, main = "Componente Irregular",
xlab = "Años/Meses", ylab = "Factores Irregulares")
Descomposición multiplicativa usando la librería stats.
library(forecast)
library(stats)
DescomposicionAditivo <- decompose(Serie.IMAE.ts, type = "multiplicative")
autoplot(DescomposicionAditivo, main = "Descomposición Multiplicativa", xlab ="Años/Meses")
Descomposición Aditiva usando la librería feasts.
library(tsibble)
library(feasts)
library(forecast)
library(ggplot2)
Yt %>% as_tsibble() %>%
model(classical_decomposition(value, type = "additive"))%>%
components()%>%
autoplot()+
labs(title = "Descomposición Clásica Aditiva, IMAE 2018-2024[Marzo]") +
xlab("Años/Meses")
Modelo multiplicativo.
- Serie temporal. \[Y_t = T_t* C_t * S_t *I_t\]
Componente Tendencia Ciclo.
library(forecast)
Ttm <- ma(Serie.IMAE.ts,12,centre = TRUE)
autoplot(object = Ttm, main = "Componente Tendencial de El Salvador 2018-2024. [Ciclo]",
xlab = "Año/Meses",ylab = "Indice")
Componente Estacional.
\[S_t = Y_t/T_t\]
SIm <- Yt/Ttm
Stm <- tapply(SIm, cycle(SIm), mean, na.rm = TRUE)
Stm <- Stm*12 / sum(Stm)
Stm <- rep(Stm, len = length(Yt))%>%
ts(start = c(2018, 1), frequency = 12)
autoplot(Stm,
main = "Componente Estacional",
xlab ="Años/Meses", ylab="Indice")
Componente Irregular.
\[I_t = Y_t/T_t* S_t\]
Itm <- Yt / (Ttm*Stm)
autoplot(Itm, main = "Componente Irregular",
xlab = "Años/Meses", ylab = "Factores Irregulares")
Descomposición multiplicativa usando la librería stats.
library(forecast)
library(stats)
DescomposicionMultiolicativa <- decompose(Serie.IMAE.ts, type = "multiplicative")
autoplot(DescomposicionMultiolicativa, main = "Descomposición Multiplicativa", xlab ="Años/Meses")
Descomposición multiplicativa usando la librería feasts.
library(forecast)
library(ggplot2)
library(feasts)
library(tsibble)
Yt %>% as_tsibble() %>%
model(classical_decomposition(value, type = "multiplicative"))%>%
components()%>%
autoplot()+
labs(title = "Descomposición Clásica Multiplicativa, IMAE 2018-2024[Marzo]") +
xlab("Años/Meses")