Abstract

The study aims to use the artificial neural network technology to classify women patients with osteoporosis and to identify the most important factors affecting the disease. Classification is considered as one of the important methods because of the necessary and repeated need to know which community among a group of communities to which a particular an observation or an individual may belong, according to the measurements or criteria obtained from observations or individuals.The research sample was concerned with women who underwent a bone density test in Baghdad Teaching Hospital for the years 2019 and 2020 and the sample size was 309 observations, 144 observations the size of the first community sample (women who underwent a bone density test and the result was not suffering from osteoporosis) and 165 observations the sample size of the second community (women who underwent a bone density test and the result was suffering from osteoporosis). These two communities represent the dependent variable where the symbol (0) was given to the first community and the symbol (1) to the second community. The number of explanatory variables was five variables, namely age and the number of years that have passed since the age menopause, weight, height and qualitative variable (was menopause early?). After inserting all the variables into the artificial neural network (the multi-layered perceptron network with feed forward) and using the ready-made program (SPSS.23), the most important results were obtained, the total correct classification rate is 78.6%, and that the two most important variables in distinguishing and classifying women with osteoporosis are: Weight variable, number of years since menopause, height, age, and qualitative variable (was the menopause early?) respectively.