In my professional field, when writing papers, I mostly use economic data at the provincial or prefectural level. Generally, these data are obtained from yearbooks, national economic and social development statistical bulletins, various statistical annual reports, or Obtained from the government website, but there are always problems with missing data caused by incomplete statistics. For example, environmental pollution-related indicators, patent application authorization numbers, etc. For these missing data, my filling methods are generally as follows:
This method is generally suitable for: unilateral missing data (usually earlier years) missing, data for recent years are generally relatively complete).
A teacher told me about the cubic spline interpolation method. The calculated missing values ??are smoother and the results may be relatively reasonable. I checked that it can be implemented in matlab, but I can only do it next time when processing data.