sgstar: Fit Seasonal Generalized Space Time Autoregressive Model

In sgstar: Seasonal Generalized Space Time Autoregressive (S-GSTAR) Model

Description

sgstar function return the parameter estimation of Seaonal Generalized Space Time Autoregressive Model by using Generalized Least Square (GLS)

Usage

sgstar(data, w, p, ps, s)

Arguments

data	A dataframe that contain timeseries data with k column as space and n rows as time.
w	a spatial weight, matrix ncol(data) * ncol(data) with diagonal = 0.
p	an autoregressive order, value must be greater than 0.
ps	an autoregressive order for seasonal, value must be greater than 0.
s	an order of the seasonal period

Value

sgstar returns output with detail are shown in the following list :

Coefficiens	coefficiens parameter model for each location
Fitted.Values	a dataframe with fit value for each location based on model
Residual	a dataframe that contain residual,that is response minus fitted values based on model
Performance	a dataframe containing the following objects:

• MSE : Mean Squared Error (MSE) for all the data combined.

• RMSE : Root Mean Squared Error (RMSE) for all the data combined.

• AIC : a Version of Akaike's Information Criterion (AIC)

• Rsquared : R^2, the ‘fraction of variance explained by the model’.

p	an autoregressive order
ps	an autoregressive order for seasonal
s	an order of the seasonal period
weight	a spatial weight
data	a dataset that used in modeling

References

Setiawan, Suhartono, and Prastuti M.(2016).S GSTAR-SUR for Seasonal Spatio Temporal Data Forecasting. Malaysian Journal Of Mathematical Sciences.10.<Corpus ID :189955959>.

Examples

library(sgstar)
data("coords")
data("simulatedata")

#create weight matrix using distance inverse matrix

z<-dist(coords,method = "euclidean")
z <- as.matrix(z)

matriksd <- 1/z
matriksd[is.infinite(matriksd)] <- 0

matriksd_w <- matriksd / rowSums(as.data.frame(matriksd))

fit <- sgstar(data = simulatedata, w = matriksd_w, p = 2,ps = 1, s =4)
fit

sgstar documentation built on May 23, 2021, 5:06 p.m.