Scholarlink

Based on statistical processing of data from the database ‘for Christ suffered’ using the principal component analysis (PCA), factual data related to repressions by region in the period 1929–1933 are analyzed. As a result of applying PCA, the dimension of the problem is reduced from 35 regions to 5 principal components (PC). The principal components by region are interpreted as waves of anti-church arrests, represented by trend regions, during the period of collectivization. Also, based on the first two PC in the dual problem, the regions of arrests are classified by month of the period. The result of the intelligent analysis of the selected multidimensional data is presented in the form of a graph of measure of the number of arrests and a thematic map of arrests, as a geoinformation layer through a mapping service. The graph and layer were generated using the ‘NKWSystem’ hypertext system for the NIKA OODBMS. As a result of the analysis, trend regions were identified, sorted in order of decreasing significance: Moscow city, Moscow Region, Arkhangelsk Region, Vologda Region, Leningrad city. The main trend regions are Moscow city and the Moscow Region with corresponding peaks of the second wave of arrests in December 1930 and May 1931.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Construction of a Multidimensional Data Cube for a Factual Database: Feature Extraction Using the Principal Component Analysis Based on the Example of Repression Statistics by Region

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Alexander Solovyev,
  • Anna Bogacheva,
  • Vladimir Tishchenko

摘要

Based on statistical processing of data from the database ‘for Christ suffered’ using the principal component analysis (PCA), factual data related to repressions by region in the period 1929–1933 are analyzed. As a result of applying PCA, the dimension of the problem is reduced from 35 regions to 5 principal components (PC). The principal components by region are interpreted as waves of anti-church arrests, represented by trend regions, during the period of collectivization. Also, based on the first two PC in the dual problem, the regions of arrests are classified by month of the period. The result of the intelligent analysis of the selected multidimensional data is presented in the form of a graph of measure of the number of arrests and a thematic map of arrests, as a geoinformation layer through a mapping service. The graph and layer were generated using the ‘NKWSystem’ hypertext system for the NIKA OODBMS. As a result of the analysis, trend regions were identified, sorted in order of decreasing significance: Moscow city, Moscow Region, Arkhangelsk Region, Vologda Region, Leningrad city. The main trend regions are Moscow city and the Moscow Region with corresponding peaks of the second wave of arrests in December 1930 and May 1931.