Risk Assessment And Survival Prognosis Analysis Of Esophageal Squamous Cell Carcinoma Based On Swarm Intelligence Optimization Algorithm

Esophageal squamous cell carcinoma(ESCC)is a common malignant tumor,and it is the sixth largest cancer related death cause in the world.The analysis of the factors influencing the survival of ESCC is based on different angles and each has its own emphasis.This paper focuses on how to establish models and improve / mix group optimization algorithm to make ESCC have high convergence.This method can achieve high accuracy through a large number of iterations.The main research work and innovation points are described as follows:Aiming at the low 5-year survival rate and overall survival rate of postoperative patients,an ESCC risk model based on adaptive lasso and relative operating characteristic(ROC)curve is proposed.In order to screen the important variables of blood data of ESCC patients,Cox regression analysis is used to establish the model,and ROC curve is used to evaluate the model.The survival rate of INR is higher than that of PNI.According to Chi square analysis,INR is significantly correlated with the degree of differentiation,infiltration and lymph node metastasis.The risk model is evaluated by INR,differentiation,invasion and lymph node metastasis.Grade 6 has the highest survival rate and the lowest risk.To solve the problem of blood data redundancy in ESCC patients,Cox regression analysis is used to analyze the blood indexes,and the factors affecting the survival or death of patients are screened out.In order to optimize the prediction accuracy of the survival level of ESCC patients,aiming at the problems of low convergence accuracy and easy to fall into local optimum of BP prediction network,the improved adaptive salp swarm algorithm(ASSA),genetic algorithm(GA)and back propagation(BP)neural network were combined.ASSA-BP model can more effectively predict the survival time of ESCC patients,shorten the training time and improve the prediction accuracy.In order to solve the problem of low convergence accuracy of BP prediction network in ESCC patients,a comprehensive model of ESCC patients’ survival level prediction based on principal component analysis(PCA)and improved ant lion optimization(IALO)is proposed.The PCA is used to reduce the data dimension and reduce the data redundancy.The main idea of IALO is operator cross mutation,which is part of differential evolution algorithm.The mutation operator is introduced into the ALO algorithm,which can enhance the diversity of population,improve the global search ability of alo algorithm,and avoid the alo algorithm from falling into local optimal.Thus,PCA-IALO-BP model is established to improve the accuracy of predicting survival level of ESCC patients.In view of the instability of blood data and training error of BP neural network in ESCC patients,a Kaplan-Meier-gray wolf optimization algorithm-BP neural network model(K-M-GWO-BP)is proposed.The purpose of this model is to reduce the dimension of data and improve the accuracy and stability of BP neural network prediction model.K-M analysis is used to screen the blood factors which are highly related to the survival level of patients,and simplify the network structure.Based on this framework,GWO optimizes the weight and threshold of BP neural network,and proposes a K-M-GWO-BP neural network model,which improves the prediction accuracy of survival level of ESCC patients.Based on the blood data of ESCC patients,this paper constructs the risk model of ESCC based on adaptive Lasso and ROC curves.Based on the single variable analysis and ASSA BP neural network,the survival period of ESCC patients is predicted.Based on principal component analysis and improved ant lion optimization algorithm(IALO),the comprehensive model of survival prediction of ESCC patients is optimized by BP neural network.Based on K-M survival analysis and GWO,BP neural network is optimized to predict the survival level of postoperative ESCC.The software is used to predict the survival level of patients with ESCC,and swarm intelligence optimization algorithm is used to optimize the prediction accuracy.