Introduction/Background Endometrial cancer is the most common neoplasm in the female genital tract in Taiwan. The aim of this study was to develop a machine learning-based classification model to predict risk factors of recurrent endometrial cancer.

Methodology A total of 13,324 patients were used to verify the feasibility and effectiveness of the treatment using data from three hospital tumor registries. Additionally, five machine learning approaches were used to develop prediction models, including LADT (Logical Analysis of Data Trees), NBT (Naïve Bayes Trees), RF (Random Forests), RT (Random Trees), and FT (Functional Trees).

Results The experimental results indicate that the RF model was of the highest accuracy. The results suggest that six of the most important recurrent risk factors were behavior, age, tumor metastasis, grade, surgical margins, and pathological stage.

Conclusion These risk factors should be monitored for early detection and the clinical features summarized in this study as additional effective treatments and appropriate interventions.