Historical Census Data Linkage: Graph-based Household Matching Method

For each developing country, historical census population data are the most useful. It provides valuable information on a certain region of a nation. The matching historical census data means that many population census data have been used to link households across the decades to support understanding the regional information of a country as gender ratio, number of population and family types, and so on. The issues with population census data consist of a lack of reliable data and many common names. The problems in household matching are linking the families and know the household structure changes during the census's years. In this paper, a graph-based data matching method with the unique Id is presented to obtain accurate household structure change results over the years as compared to the baseline graph-based household matching method. The proposed method will accomplish household matching in both single and multiple households. It is also able to understand variations in the organization of the households between the decades. Consequently, the graph-based household matching method achieves 96% accuracy that outperforms the other baseline graph-based household matching method.