IMPUTATION OF MUNICIPAL STATISTICS DATA
T.B. Skripkina
Novosibirsk State University of Economics and Management, Novosibirsk, Russian Federation
Keywords: импутация статистических данных, восстановление пропусков данных, муниципальная статистика, метод K ближайших соседей, критерий согласия Пирсона, imputation of statistical data, restoration of data omissions, municipal statistics, K nearest neighbor method, Pearson's consent criterion
Abstract
The article considers the key problem of conducting statistical analysis on an array of municipal statistics: the presence of a significant amount of missing data. The concept and types of data imputation are given. The application of the K nearest neighbor method for imputation at the stage of using official statistics data contained in the database of municipal indicators is justified. The author’s algorithm for restoring omissions in the data set of municipal statistics indicators using the STATISTICA system is proposed. Verification of the imputation results was performed by comparing the distributions of the original and restored data based on the Pearson’s consent criterion.
|
