| With the continuous improvement of Chinese economic level,China’s medical industry has made great progress.The demand for high-quality,high-level medical services is eagerly anticipated,and the country is widely calling for the strengthening of artificial intelligence applications in healthcare.Although the current information technology construction in hospitals has begun to bear fruit,there are still some problems,which are the lacking of maintenance for medical tools,cumbersome usage,and Inadequate intelligence.In this thesis,a model for predicting survival of colorectal cancer patients based on the XGBoost method and using the SEER dataset is investigated,and the model is applied to the evaluation of colorectal cancer tumor treatment options.At present,survival prediction of tumor patients mainly relies on patients’ health information and tumor-related information where there are many related studies.However,most of them ignore the impact of the treatment method taken by patients on survival.In this paper,based on existing studies,the treatment-related attributes in the SEER dataset were introduced into the experiment,and the data preprocessing was completed by data transformation,null filling,unique heat coding,normalization,etc.The XGBoost method was improved to obtain a survival prediction model for colorectal cancer patients with an accuracy of 88%and AUC of 75%in the validation set,which can be integrated and used in an assisted decision-making system.The system is based on the actual workflow,specific business scenarios of medical personnel and three clinical application scenarios,patient treatment,patient review and patient data management,are chosen to analyze,design and implement a fully functional,beautiful and easy-to-use colorectal cancer assisted decision-making system,which can assist doctors in making decisions on treatment plans and patient status assessment during review and complete some It is of practical value and relevance. |
PDF Full Text Request