Hypothyroidism diagnosed cases using big data analysis with k-means clustering method

In the past, it was difficult to make big data explainable. Until recently, it was not commonly used to cause huge costs and long-term period. Nowadays, it is possible to analyze big data with hardware and software developments and despite the expanding data repository. Aim of this thesis is to analyze congenital hypothyroidism, hypothyroidism and acute thyroiditis diagnosed cases' laboratory measurements and socio-demographic characteristics using big data. In this thesis, were included 21125 patients to the study by taking into consideration the first diagnosis' laboratory measurement values which were between 2005 and 2018 accessible defined as above diagnosed cases from the database of Bursa Uludag University Health Application and Research Center. The data were divided into two clusters according to laboratory measurement and demographic variables. In addition to the usage of big data, clusters were analyzed with the Effect Size of Cliff's Delta. Reference values to diagnose for laboratory measurements which are Free T3 and Free T4 were compatible with results of the big data analysis in our study, although TSH laboratory values were incompatible. The differences after analyzing big data should be considered to evaluate and investigate with planned and controlled studies.