Table 1: Model specifications for consumer price inflation divisions
"Divisions, selected ARIMA order and seasonal component"
,Model specification,
Division,"Selected ARIMA Order (p,d,q)","Seasonal Component (P,D,Q)"
Food and non-alcoholic beverages,"(1,1,1)","(0,0,1)"
Alcohol and Tobacco,"(0,1,1)","(1,0,0)"
Clothing and footwear,"(0,1,0)","(1,0,0)"
Housing and household services,"CPIH (1,1,0)","(0,0,1)"
,"CPI (0,2,1)",
Furniture and household goods,"(0,1,1)","(0,0,1)"
Health,"(1,1,0)","(1,0,0)"
Transport,"(0,1,2)","(1,0,0)"
Communications,"(0,1,1)","(0,0,1)"
Recreation and Culture,"(3,1,1)","(1,0,1)"
Education,N/A,N/A
Restaurants and hotels,"(0,1,0)","(0,0,1)"
Miscellaneous Goods and Services,"(0,2,1)","(1,0,1)"
Source: ,Office for National Statistics
Notes
1,"An ARIMA model specification is given in terms of three components, typically denoted (p, d, q). The p component is used to denote the number of time lags of the autoregressive model; that is, the number of previous periods for which the previous values of the variable are used as predictors of the current value. The d component is the minimum number of times that we need to take the previous value of the series from the current value in order to make the series stationary (independent of time). This is so we can apply the autoregressive and moving average components of the model. The q component is the number of previous periods used within the moving average calculation."
2,"We separate the non-seasonal from the seasonal components of the model. The latter consider each component with respect to the seasonality of the data, and they are specified separately. For example, how many times the series must be differenced with respect to the same time in the previous season."