|
gptkbp:instanceOf
|
statistical analysis
|
|
gptkbp:alternativeTo
|
gptkb:exponential_smoothing
machine learning models
state space models
|
|
gptkbp:appliesTo
|
economics
finance
weather forecasting
sales forecasting
|
|
gptkbp:assumes
|
linearity
no autocorrelation in residuals
|
|
gptkbp:canBe
|
non-seasonal
seasonal (SARIMA)
|
|
gptkbp:category
|
univariate time series models
|
|
gptkbp:component
|
autoregressive (AR) part
integrated (I) part
moving average (MA) part
|
|
gptkbp:diagnostics
|
gptkb:Ljung-Box_test
ACF plots
PACF plots
residual analysis
|
|
gptkbp:extendsTo
|
gptkb:SARIMA
gptkb:ARIMAX
|
|
gptkbp:fullName
|
gptkb:Autoregressive_Integrated_Moving_Average_models
|
|
gptkbp:hasModel
|
non-stationary time series
stationary time series
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
ARIMA models
|
|
gptkbp:implementedIn
|
gptkb:Python
gptkb:SAS
gptkb:MATLAB
R
|
|
gptkbp:inferenceMethod
|
maximum likelihood
least squares
|
|
gptkbp:introduced
|
gptkb:George_Box
gptkb:Gwilym_Jenkins
|
|
gptkbp:introducedIn
|
1970
|
|
gptkbp:limitation
|
sensitive to outliers
cannot model non-linear relationships
requires large data
|
|
gptkbp:parameter
|
d (degree of differencing)
p (order of AR part)
q (order of MA part)
|
|
gptkbp:publishedIn
|
gptkb:Time_Series_Analysis:_Forecasting_and_Control
|
|
gptkbp:relatedTo
|
gptkb:Kalman_filter
gptkb:Box–Jenkins_methodology
gptkb:vector_autoregression_(VAR)
|
|
gptkbp:requires
|
parameter estimation
stationarity (after differencing)
|
|
gptkbp:usedFor
|
time series forecasting
|
|
gptkbp:bfsParent
|
gptkb:SPSS_Forecasting
|
|
gptkbp:bfsLayer
|
6
|