Decision tree based sub-healthy states detection system

Large number of people are under Sub-healthy state now, which means they are half healthy and half sick. And this state is tough to detect and treat for undergoing western medicine methods for its non-organic diseases characteristics. However, because of the holistic and dialectical view of Traditional Chinese Medicine (TCM), it´s generally more effect for treating the sub-healthy state by TCM solutions. As one of the most import methods in TCM, wrist pulse diagnosis approach can be used to handle the detection of sub-healthy state. We propose a sub-healthy states detection system by building a decision tree based on wrist pulse data. To preprocess the raw data, we first develop a preprocessing model and then build the decision tree on both time domain features and frequency domain features. Not only high accuracy rates but also good recognition rates are achieved in our experiments.