Study On The Syndrome Differentiation Model Of Colorectal Cancer Based On Integrated Learning

In recent years,the incidence rate and mortality rate of colorectal cancer have been high.Colorectal cancer has become one of the malignant tumors which threaten our health seriously worldwide.At present,the conventional treatment will bring many adverse reactions to patients and affect the quality of life after treatment.Many studies have shown that traditional Chinese medicine has unique advantages in the prevention and treatment of colorectal cancer,which can effectively prolong the survival of patients.Syndrome differentiation is the core and key of traditional Chinese medicine in the treatment of colorectal cancer,but due to the lack of objective dialectical evaluation indicators,the dialectical results have great uncertainty,which greatly reduces the effect of traditional Chinese medicine in the diagnosis of colorectal cancer.Machine learning technology can find the internal relationship between symptoms and syndrome types.By building an intelligent dialectical classification model,it can assist clinicians to make decisions and promote the process of objectification of traditional Chinese medicine.Considering that the integrated learning method has better generalization ability and higher prediction accuracy,this study uses the integrated learning algorithm to do the dialectical typing research for the collected colorectal cancer data.First of all,the colorectal cancer cases included in the study were sorted out and unified according to the corresponding standards and guidelines,and the general situation of patients' gender,syndrome type and so on were statistically analyzed,and the distribution results of symptoms and signs,tongue image and pulse as well as the main symptom distribution of each of the six syndrome types were summarized and analyzed.Then,the importance of each symptom category is calculated by using infogainattributeeval evaluator,and the main symptoms of colorectal cancer arescreened out,and the classification accuracy of each model before and after feature optimization is evaluated.The results show that after symptom feature optimization,the prediction accuracy of each machine learning model is greatly improved,which verifies the necessity of feature selection before model construction.Finally,Comparing the classification performance of the models through indicators such as kappa coefficient,and the accuracy of the integrated learning model is 93.55%,and the kappa coefficient is 0.9202,which is superior to the other three classical models.It shows that the integrated learning method proposed in this study has superior performance in the dialectical classification of colorectal cancer in traditional Chinese medicine.This method can be preferred to be used for the intelligence of colorectal cancer syndrome types to distinguish.