Prognostic Model Analysis And Early Warning System Design Of Small Cell Lung Cancer Based On Improved Bayesian Network

Lung cancer is a kind of the commonest cancers in the world,and it also seriously threatens our human health.Many patients are already in the advanced stage once they are discovered,but the problem of how to intervene and improve the overall remission proportion and prognostic survival proportion of lung cancer patients has not yet been resolved.According to histological classification,small cell lung cancer(SCLC)is classified as lung cancer.It has a very high value-added rate,easy to produce early metastasis and poor prognosis.Therefore,the prognostic analysis of the diseased population is of positive significance.At present,big data,machine learning,and picture dealing have played a positive role in cancer diagnosis and postoperative prediction.In view of the existing literature,this thesis understands the status quo and development trend of small cell lung cancer,and improves the structure learning of Bayesian network through ant colony algorithm,and better probes the major factors that influence the prognosis of SCLC and danger estimation.A model was built on the selected data set,and an early warning system was designed.The detailed work as follows:First,based on the improved Bayesian network,a model for prognostic survival hazard dissection of SCLC was constructed.Here,ant colony algorithm and Bayesian network are combined.This thesis first reduces the ant colony search space and simplifies the model complexity through the conditional independence in the Bayesian network;then uses a new heuristic function of the ant colony algorithm to cut down the possibility of local optimal issues and obtains the improved Bayesian network.Secondly,this thesis selects a small cell lung cancer prognostic data set from the US SEER database and performs a series of preprocessing such as clustering dimensionality reduction,coding,and missing value interpolation.The risk variables that are highly related to the prognosis of SCLC are screened out,and clinical data that can be more suitable for the model are obtained.By comparing with other prediction models introduced in the thesis,experiments show that the model method in this thesis has better prediction performance.Lastly,in view of the SCLC prognosis model constructed in this thesis,designs an early warning system for the prognosis of SCLC,it not only meets the needs of cancer patients' prognostic survival risk estimation,but also helps doctors adjust treatment plans in a timely manner with reference to the predicted results and promotes the implementation of personalized treatments for the prognosis of small cell lung cancer.