Comparing the Effectiveness of Two Simple Decision Tree and Random Forest Algorithms in Predicting Osteoporosis in Active Middle-Aged Men

Osteoporosis is called a “silent disease” because it has no symptoms until a bone is fractured. Therefore, its early detection before occurrence of fracture is important. Using data mining algorithms can help access the information hidden in the data. This study aims to compare two simple decision tree and random forest algorithms to predict osteoporosis in active middle-aged men.

A total of 256 middle-aged men referred to Ayatollah Kashani Hospital in Tehran, Iran during 2017-2020 participated in this study. Data analysis was carried out in MATLAB software version 2020.  Evaluation was performed using the confusion matrix and based on accuracy and precision criteria.

Out of 103 factors related to personal information, lifestyle, and disease, 11 were selected as inputs to the algorithms. The results showed that the random forest algorithm had a better performance (73.4% accuracy and 68.07% precision) compared to simple decision tree.

The data mining algorithms can be effective in predicting osteoporosis in active middle-aged men. These algorithms can be used for early treatment and rapid diagnosis of osteoporosis and prevent the occurrence of bone fractures and their irreparable complications.