Now that we have extracted the trend, we can remove it from the initial time series. This is done by division. We will be left with the product of the seasonal and noise components. Thus, SN = Y/T.
To calculate the fraction between TimeArray objects, we'll use the element-wise division operator, ./:
julia> sn = unemployment_data ./ moving_avg
145×1 TimeArray{Float64,2,Date,Array{Float64,2}} 2005-12-01 to 2017-12-01
We will get the following result:
Plotting the resulting TimeArray will give us a clearer image of the product of the seasonality and noise components:
julia> plot(sn)
This results in the following plot: ...
