Clustering with Prototype Extraction for Census Data Analysis

TL;DR

This paper introduces a novel clustering technique for analyzing census data to uncover hidden societal patterns and factors influencing individual decision-making, demonstrated through an experimental case study.

Contribution

The paper presents a new clustering-based method for identifying influential factors in census data, focusing on decision-making processes like childbirth.

Findings

Clusters reveal key factors influencing behavior

Prototypes help distinguish between different respondent groups

Method successfully applied in a census data case study

Abstract

Not long ago primary census data became available to publicity. It opened qualitatively new perspectives not only for researchers in demography and sociology, but also for those people, who somehow face processes occurring in society. In this paper authors propose using Data Mining methods for searching hidden patterns in census data. A novel clustering-based technique is described as well. It allows determining factors which influence people behavior, in particular decision-making process (as an example, a decision whether to have a baby or not). Proposed technique is based on clustering a set of respondents, for whom a certain event have already happened (for instance, a baby was born), and discovering clusters' prototypes from a set of respondents, for whom this event hasn't occurred yet. By means of analyzing clusters' and their prototypes' characteristics it is possible to identify which factors influence the decision-making process. Authors also provide an experimental example of the described approach usage.