| Summary: | 碩士 === 國立中興大學 === 資訊管理學系所 === 106 === In recent years, machine learning has been widely used in medical diagnosis. For example, it has good efficacy in predicting tumors and cancer. At present, some thesis have proposed effects on rheumatoid arthritis and periodontal disease, but due to no infectious agent has been consistently linked with rheumatoid arthritis and there is no evidence of disease clustering to indicate its infectious cause, periodontitis has been consistently associated with rheumatoid arthritis. In this thesis, we use Ensemble methods including random forest, bagging, adaboost, gradient boosting to predict the positive correlation between the periodontitis and rheumatoid arthritis, and further the prediction the efficacy of rheumatoid arthritis, which features belong to periodontal disease will affect its accuracy. The experimental results show that using machine learning to find out the features such as gender, smoking, bleeding on Probing can confirm by other thesis. It has a positive correlation with rheumatoid arthritis and periodontal disease. Then, when predicting the efficacy of rheumatoid arthritis, there is an accuracy of more than 80%. When the relevant features of periodontal disease delete, the accuracy rate reduces to 40%. It can show that periodontal disease has a great impact in Rheumatoid arthritis. We hope to use this method to improve medical accuracy. This thesis proves to you that machine learning not only helps doctors increase the accuracy of predictions and reduces the misjudgment as possible, but also helps us provide for more convenient and better quality of medical life.
|
|---|
