Temporal Disaggregation

3 comments:

  1. Hi Shishir,
    I have gone through through your ARIMA blog.It is awesome and worth reading.
    I want you to please through some light on Temporal disaggregation as well.
    As right now i am looking for it to convert my bimonthly sales data to monthly.
    Your help would highly appreciate..Eagerly waiting for it :)

    ReplyDelete
  2. Try Check this matlab routines:
    https://www.mathworks.com/matlabcentral/fileexchange/39770-temporal-disaggregation-library
    Try see tempdisagg in R:
    https://cran.r-project.org/web/packages/tempdisagg/tempdisagg.pdf

    ReplyDelete
  3. If you want a quick quaterization here is quick python code.:

    def year2quarter(x,method):
    x = np.matrix(x)
    ny = x.shape[0]
    nv = x.shape[1]
    nq = ny * 4
    y = np.zeros((nq,nv))
    for i in range (0,ny-1):
    for j in range (0,4):
    d = (x[i+1,:]-x[i,:])/4
    y[4*i+j,:]=x[i,:]+j*d
    # treatment of the first year values if they are zeros
    for k in range (0,nv):
    if x[0,k]==0:
    y[0:4,k]=0
    # treatment of the last year
    if method == 1:
    d = (np.zeros((1,nv))-x[ny-1,:])/4
    for j in range (0,4):
    y[4*(ny-1)+j,:]=x[ny-1,:]+j*d
    elif method == 2:
    for j in range (0,4):
    y[4*(ny-1)+j,:]=x[ny-1,:]
    return y

    ReplyDelete