Automation of predicting the risk of osteoporosis

Purpose of research . To increase the accuracy of forecasting by developing optimal predictive models that increase the reliability of the identified estimates in the diagnosis of osteoporosis using an expert system. Methods. In this work, parametric statistical methods were used to confirm the validity of the assumption of the normal distribution of all studied sample populations. Based on the results obtained, correlation and regression analyses are considered as the best approach for studying interrelations, for developing predictive models with determining the confidence interval of prediction and comparing the range of the final indicator with the probability of disease occurrence. Result. During the study, predictive models were constructed for each identified patient group along with limits of confidence intervals for predicted values of key parameters. Regularities in the mutual dependence between specific target factor ranges and degrees of pathology progression probabilities have been established. An expert information system has been developed, implementing a complete informational support system designed to identify risks of diseases development in patients based on provided data regarding chest tissue density. Conclusion. The developed information system enables predicting the likelihood of illness for patients (both men and women) whose age exceeds the specified range (10–70 years); however, the reliability of such predictions remains ambiguous since the prognostic models underlying the system's operation are built upon data from individuals aged within the period of approximately 10 to 72 years.