The objective of this study is to cluster 40 different departments at the Adnan Menderes University Hospital according to some variables. The data used in the study was obtained from the 2014 statistics of the Adnan Menderes University. Among the hierarchical clustering approaches, complete link method was used as the study attempted to determine the way of merging or partitioning of clusters. The study determined that all independent variables had a significant effect in clustering as the result of the ANOVA test which was done for clustering purpose. Chi-square means and discriminant analysis method was used in order to provide evidence on the validity of the cluster results obtained in the analysis. The results of the study were discussed in two stages. The first stage included the analysis of all 40 units while variables of relevant expense, package loss and SGK deduction were evaluated. The second stage included the analysis of 37 units while the variables of expense, lecturer, assistant, nurse, number of personnel, number of polyclinic rooms, polyclinic area, number of service beds and service area were evaluated. Upon the analysis in the first stage, it was determined the units were gathered under 5 clusters. The analysis showed that both the orthopaedics and oncology units were a cluster on their own while the units of hematology and brain surgery were included in the same cluster. The fourth clusters consist of the units of Cardiovascular Surgery, General Surgery, Emergency and Cardiology while the fifth cluster consists of the all other units. As a result of the analysis in the second stage, it was observed that the number of clusters and units within clusters didn’t vary. In order to determine the validity of the results of the study, it was determined that the number of clusters obtained by calculating Wilk’s lambda coefficient was the same with the number of clusters determined by the complete link method. According to the findings obtained in the study, it was determined that the units with highest expenses made a single unit and it is believed that the expenses of both units were significantly different from other units.

Clustering, discriminant analysis, ANOVA, hospital, clinics

[1] Mannila, H. Data mining: Machine learning, statistics, anddatabases. Proc. Of the 8th International Conference on Scientificand Statistical Database Management, 1996, 1-6
[2] Fayyad, U. M., G. Piatetsky-Shapiro, and P. Smyth. Fromdataminingtoknowledge discovery in databases. AI Magazine,1996, 37-54.
[3] Yünel Y. K-means Kümeleme Algoritmasının Genetik Algoritma Kullanılarak Geliştirilmesi, p. 1-2.Zeytinoğlu, F. Ç. (2014) Kümeleme Analizi: Kültür İstatistiklerine Göre İllerin Sınıflandırılmasına Yönelik Bir Çalışma, İstanbul Ticaret Üniversitesi Sosyal Bilimleri Dergisi,2010, 13 (25), 301-320.
[4] Çokluk, Ö, Şekercioğlu, G. ve Büyüköztürk, Ş. Sosyal Bilimler İçin Çok Değişkenli İstatistik SPSS ve Lisrel Uygulamaları, 2014,Ankara: Pegem Akademi.
[5] Kaya, V. ve Türkmen, A. Küresel Krizin Üst Orta Gelir Gurubu Ülkelere Makro Ekonomik Yansımaları, Atatürk Üniversitesi İktisadi ve İdari Bilimler Dergisi, 2013, 27 (4), 317-338
[6] Neil, T.H. AppliedMultivariate Analysis,Springer-Verlag, New York, 2002
[7] Kalaycı, Ş. SPSS Uygulamalı Çok Değişkenli İstatistik, Ankara: Asil Yayın Dağıtım, 2014
[8] Alpar, R. Uygulamalı Çok Değişkenli İstatistiksel Yöntemler, Ankara: Detay Yayıncılık, 2011
[9] Duran, B.S. ve Odell, P.L. Cluster analysis: lecturenotes in economicsandmathematicalsystems, Econometrics. New York: Springer-Verlag.1974
[10] Aldenderfer, M. S. ve Blashfield, R. K. Cluster Analysis, Beverly Hills: SagePublications, 1984
[11] Zeytinoglu, I. U.,Denton, M., Davies, S., Baumann, A., Blyte, J. ve Boos, L.Retainingnurses in theirhospitalsand in theprofession: Effects of jobpreference, unpaidovertime, importance of earningsandstress. HealthPolicy, 2006,79, 57-72.
[12] Gürsoy, U.T. ve Şimşek, G. Veri Madenciliği ve Bilgi Keşfi. Ankara: PegemAkademi, 2009
[13] Anderberg, M. R. Cluster Analysis forapplications, New York: AcademicPress, 1973
[14] Özdamar, K. Paket Programlar ile İstatistiksel Veri Analizi. Eskişehir: Kaan Kitabevi, 2004
[15] Milligan, G. W. ve Cooper, M. C. A Study of theComparability of ExternalCriteriaforHierarchical Cluster Analysis,MultivariateBehavioralResearch, 1986, 21,441-458.
[16] Tatlıdil, H. Uygulamalı Çok Değişkenli İstatistiksel Analiz, H.Ü. Fen Fakültesi İstatistik Bölümü, Ankara, 1992
[17] Mardia, K. V., Kent, J. T. ve Bibby, J. M. Multivariate Analysis, London: AcademicPress, 1989.
[18] Çakmak, Z. Kümeleme Analizinde Geçerlik Problemi ve Kümeleme Sonuçlarının Değerlendirilmesi, Dumlupınar Üniversitesi Sosyal Bilimler Dergisi, 1999,3, 187- 205.
[19] Aslan, Ş., Özata, M. ve Atayeter, C. Sağlık İşletmelerinde Ekip yönetimi: Fırsatlar ve Sınırlılıklar,Standard Ekonomik ve Teknik Dergi,2004, 43 (516), 17-23.
[20] Frawley, W.J. Piatetsky-Shapiro, G. Matheus, C.J. Knowledge discovery in databases, AAAI/MIT Press, 1991, pp. 1–27
[21] Larose, D. T. k-NearestNeighborAlgorithmknowledge in Data: An Introduction to DataMining, ABD: Wiley Online Library, 2005
[22] Weiss, S. M.andKulikowski, C. A. Computersystemsthatlearn. San Mateo, CA: Morgan Kaufmann,1991
[23] Poulsen, J. ve French, A. Discriminantfunctionanalysis (DA).Retrieved August152015, fromhttp://userwww.sfsu.edu/efc/classes/biol710/discrim/discrim.pdf, 2003
[24] Aaker,D. A, Kumar, V. andDay, S. G. Marketing Research, Fifth Edition, Canada, John WileySons, Inc. 1997.
[25] Klecka, W. Discriminant Analysis. London: Sage Publications, 1980.
[26] Djomou, Z. Y.,Monkam, D. ve Woafo, P. Variabilityandtrends of local/regional scale surface climate in northern Af ricaduring the twentieth century, Theoretical and Applied Climatolog, 2013,117, 625-641.