Time Series Analysis On Consumption Of Electricity In Kwara State
[A CASE STUDY OFFA NEPA DISTRICT OFFICE FROM 2001-2015]

2.9    TIME SERIES MODEL
    There are types by which the involvement or component interests to give rise to time series.
    The multiplicative model
    The additive model
MULTIPLICATIVE MODEL
Y_t=S_t×T_t×C_t×I_t
The steps are
    Compute T_t (using method of least square method), since this will keep both C_t×I_t intact in the series and avoid unnecessary mixing up of value of data at the beginning and at the end.
    Compute S.V using Y_t=S_t×T_t×C_t×I_t    
    Compute Y_t/(T_t S_t )
    Apply an order 3 or 4 M.A. method (it is hope that during process, it would be removed).
    The result (iv) is the cyclical variation.
ADDITIVE MODELS
    The additive models of the time series i.e.
Y_t=S_t+T_t+C_t+I_t
    This leads to (by interchanging of variable)
S_t=Y_t-T_t-(C_t+I_t )  or S_t=Y_t-T_t- Residual value
    The steps are
    Calculate T_t (trend) value using suitable M.A. method
    Subtract T_i value from Y (observed value)
    Find quarterly total and mean after arranging the M.A in their respective periods.
    We have unadjusted seasonal variation (S.V) to be adjusted
    Find the grand mean if the sum of the mean S.V in (iv) is not equal to zero.
    Get the absolute value of /Q.M – G.M/
    The result in (vi) are the adjusted S.V for each group which sum must be equal to zero.