Pattern mining and machine learning for demographic sequences

Abstract

In this paper, we present the results of our first studies in application of pattern mining and machine learning techniques to analysis of demographic sequences in Russia based on data of 11 generations from 1930 to 1984. The main goal is not prediction and data mining methods themselves but rather extraction of interesting patterns and knowledge acquisition from substantial datasets of demographic data. We use decision trees as techniques for demographic events prediction and emergent patterns for searching significant and potentially useful sequences.