Determining the factors related to osteoporosis by supervised machine learning methods

In this study, it was aimed to examine the relationship between demographic and biochemical parameters and bone mineral density of patients aged 50 and over who applied to Fırat University Medical Faculty Hospital by using supervised machine learning methods. The population of the retrospective study consisted of all individuals aged 50 and over who applied to Fırat University Medical Faculty Hospital between January 2018-January 2022. The whole universe was reached without making a sample calculation. The research data were obtained by the researcher by scanning the patient files with a retrospective questionnaire in which the age, gender, chronic diseases and biochemical parameters of the participants were scanned. Bone densitometry scores of all participants were recorded and participants were grouped according to the t-score result. Of the 470 participants included in the study, 15,3% were male and 84,7% were female. The median age of the participants was 66,0 (50,0-94,0) years. Osteopenia and osteoporosis were statistically higher in women (p=0,001). In the supervised machine learning method, Jrip algorithm has 2 rules and PART algorithm has 65 rules. The metric values of the algorithms were found to be 49,36% and 94,02%, respectively. As a result; The PART algorithm can predict osteoporosis-related factors with high accuracy. In this context, we believe that new and large-scale research is needed for the use of artificial intelligence applications in the field of medicine.