Osteoporosis has recently been acknowledged as a major public health issue in developed countries because of the decrease in the quality of life of the affected person and the increase in public costs due to complete or partial physical disability. The aim of this study was to use the J48 algorithm as a classification task for data from women exhibiting changes in bone densitometry. The study population included all patients treated at the diagnostic center for bone densitometry since 2010. Census sample data collection was conducted as all elements of the population were included in the sample. The service in question provides care to patients via the Brazilian Unified Health System and private plans. The results of the classification task were analyzed using the J48 algorithm, and among the dichotomized variables associated with a diagnosis of osteoporosis, the mean accuracy was 74.0 (95% confidence interval [CI], 61.0-68.0) and the mean area under the curve of the receiver operating characteristic (ROC) curve was 0.65 (95% CI, 0.64-0.66), with a mean sensitivity of 76.0 (95% CI, 76.0-76.0) and a mean specificity of 48.0 (95% CI, 46.0-49.0). The analyzed results showed higher values of sensitivity, accuracy, and curve of the ROC area in experiments conducted with individuals with osteoporosis. Most of the generated rules were consistent with the literature, and the few differences might serve as hypotheses for further studies.

Background: Bayesian classifiers have the advantage of determining the class to which a given record belongs compared to traditional classifiers, taking as base the probability of an element belonging to a class. Thus, the diagnosis of diseases such as osteoporosis and osteopenia can become faster and precise. This paper presents an assessment of the accuracy of the Bayesian classifiers Bayes Net, Naive Bayes and Averaged One-Dependence Estimators to support diagnoses of osteopenia and osteoporosis. Method: The methodology that guided the development of this research relied on the choice of database, the study of the Bayes Net, Naive Bayes and Averaged One-Dependence Estimators algorithms, and the description of the experiments. Results: The algorithm with the highest specificity was Bayes Net, (53.0±0.27). The highest accuracy was obtained using the AODE classifier (83.0±0.17). Our results showed higher mean instances correctly classified using the Naive Bayes algorithm (82.84±14.42), and the average of incorrectly classified instances was higher for Bayes Net (17.46±14.76). Conclusion: Based on the statistical measures analyzed in the experiments (instances classified correctly and incorrectly, the kappa coefficient, mean absolute error, sensitivity, specificity, accuracy, recall, F-measure, and Area Under Curve (AUC)), all classifiers showed good results, thus, given these data, it is possible to produce a reasonably accurate estimate of the diagnosis.