A Novel Combined Approach of k-Means and Genetic Algorithm to Cluster Cultural Goods in Household Budget

摘要

Setting up household spending and leading to an efficient and optimal usage is one of the important issues that every family faces. In this paper, we consider a sample of 35,000 households, among them we evaluate features of households that allocated a larger share of the budget to cultural goods and cluster them to extract common social and economic characteristics using Combination of k-means and genetic algorithm. GA as a meta-heuristic optimization algorithm increases the speed of achieving optimal solutions in k-means algorithm. The families’ priorities to spend their budget in rural and urban areas show that in most of the families with a high level portion of cultural goods, food and drinks, smokes, and education are three categories which have a higher priority than other groups. The results show the highest accuracy that there are three well separated and compact clusters, which for the fitness are accredited by Davies–Bouldin index by calculating inter and intra distances.