A Statical approach on determining the homogenous province groups on health in Türkiye

ABSTRACT Master of Science Thesis A STATISTICAL APPROACH ON DETERMINING THE HOMOGENOUS PROVINCE GROUPS ON HEALTH IN TÜRKİYE BAHATTİN HAMARAT Anadolu University Statistical Basic Science Branch of Science Institute Supervisor: Prof. A. Fuat YÜZER 1998, Page 75 It is necessary to determine health policies to be applied and precautions to be taken for the growth on health to be in balance in Türkiye. For this reason the relation between the provinces which are similar themselves on the distribution of investments and on human and physical sources should be established. Then a balance should be formed between the investments to be made for the provinces according to this relation. Then the provinces which have not been developed on health and the provinces which are developing may be risen to the level of developed provinces. In this study determining the homogenous province groups bearing the same features which will provide analytical and visual information for the processes for forming regional development policies in the field of health has been aimed. In this study grouping the provinces in Türkiye according to their health, population and some economical and cultural indicators has been made. The 30 variables in indicators have been studied and the clustering analysis has been used for obtaining the province groups. The different clustering techniques have been applied according to different distance measures one by one and it has been determined that Ward Clustering Method gave the most appropriate result according to Squared Euclidean Distance. As a result the provinces have been divided into 8 subsets. In the stage whether the provinces have been appointed into the groups incorrectly or not, discriminant analysis for the results of clustering analysis has been applied and the result of clustering the provinces successfully has been achieved. And then it has been determined that the variables have important parts in discrimination. Keywords: Clustering Analysis, Discriminat Analysis, The Health Level of Provinces, Similarity, Dissimilarity.