Parameter weighted distance clustering for panel data – taking urban real estate development as an example

Abstract:
Considering that panel data contain both cross-sectional data and time series data, we reconstruct the similarity measure function, and propose the parameter weighted distance clustering method. The feature weight is measured by entropy weight method, and the similarity measure function is reconstructed from multiple angles, which not only fully extracts panel data information, but also can be directly applied to panel data with inconsistent data length and missing data. An empirical study is carried out in urban real estate development, and the results show that this method has good application in solving panel data clustering.