sgstar {sgstar}

Fit Seasonal Generalized Space Time Autoregressive 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