Clustering According to Cultural Structures of Cities in Turkey based on Fuzzy C-means Method
Nihat Tak · RePEc: Research Papers in Economics
The study investigates the cities of Turkey regarding to their cultural structures by using K-means, Fuzzy C means algorithms and Model Based clustering method. According to K-means algorithm, we have two clusters and there are three cities in the first cluster, which are basically the first three biggest cities in Turkey. According to Fuzzy C means method, Istanbul is itself one cluster, Izmir, Ankara, Antalya are in one cluster, 28 cities are in another cluster, and the rest (49) of the cities are in the same cluster. Fuzzy C-means clustering method with 4 clusters is more informative than K-means method with 2 clusters. According to Model Based clustering, we had 29 cities in the first cluster and 52 cities in the second cluster. However, grouping Istanbul in the same cluster with small city like Van may be mistake. Therefore, Model Based Clustering might not be the best method for this data set.