HİYERARŞİK KLUSTER (KÜME)TEKNİĞİ KULLANILARAK TÜRKİYE’DE İLLERİN SOSYO-EKONOMİK BENZERLİKLERİNİN ANALİZİ

Abstract

In this study, 54 socio-economic variables were used to determine groups for Turkish provinces which represent similar characteristics. To achive this goal hierarchical cluster analyses was decided to be an appropirate statistical test. While doing determination of number of clusters for 81 provinces, the data set was tested with 7,10 and 15 clusters(groups). As a result of testing clustering processes, the group with 15 members was selected to use during analyses. The analyses were supported with dendogram and agglomerative schedule. Squared oklid distance and pearson proximity matrix were used to calculate distances between different provinces to find out province groups which have similar socio-economic character. The results indicate that, the most similar provinces are Bitlis and Mardin, and the least similar provinces are İstanbul and Kars. Consequently, this study is able to determine social and economical differences of Turkey’s provinces with aid of hierarchical cluster technique

Keywords

• hierarchical cluster analysis
• pearson proksimity matrix
• turkey
• provinces
• socio-economic regions