A Trustworthy Decision-Making Algorithm Based on Vote Trees of Random Forest for Abdominal Aortic Aneurysm Diagnostic

In order to reduce the abdominal aortic aneurysm (AAA) mortality and avoid surgery and the economic burden caused by this disease, it is important to predict the probability of AAA to take early intervention and control. Recently, many researches have focused on the analysis of AAA after aneurysm forms, and obviously, this posterior intervention has already caused great harm to the patient's body. Effective intervention and prevention methods for early AAA are lacking. Therefore, this article proposes a decision-making algorithm based on vote trees of random forest, and it is used to diagnose early AAA according to the features from our samples. The graphical vote trees algorithm shows the decision-making path of intelligence (AI) system. It can quickly help doctors to understand the whole process of AI decision-making. Experiments indicate that our model can detect the probability of AAA with higher accuracy than other algorithms, and the results are trustworthy and stable for each decision.