Aid decision of Chinese traditional patent medicine based on manifold ranking

Chinese Patent Medicine is widely used to treat many disease in China since better efficacy can be achieved in the clinical practice. However, because of lack of Traditional Chinese Medicine (TCM) knowledge, many western physicians are confusing with the utility of Chinese Patent Medicine, which are exactly the main force of clinical prescriptions. Therefore, the aid decision method is urgent and necessary to help the western physicians rationally use Chinese Patent Medicine. In this paper, Manifold Ranking (MR) is employed to developing the aid decision model of Chinese Patent Medicine. There are two stages in the aid decision model Firstly, the underlying relation between the symptoms and the Chinese Patent Medicines is obtained based on MR during the learning process. Secondly, given the symptoms of a new patient, the aid decision of Chinese Patent Medicine is able to be given during the decision process. 115 patients with stroke are taken as the examples of training and testing the aid decision models. The experimental results reveal that Chinese Patent Medicine is able to be differentiated and the high accuracy of aid decision is also obtained based on some symptoms.