Application Of Swarm Intelligence Algorithm In Parameter Optimization Of Liver Disease Detection

With the improvement of people’s living standards,health problems are becoming increasingly important.Liver disease refers to the pathological changes in the liver,which is a common disease with great harm.Liver disease is caused by excessive drinking,inhalation of harmful gases,intake of contaminated food and drugs.Liver has many basic functions.Liver disease brings many problems to the provision of medical services.The diagnostic methods of liver disease have been widely concerned by scholars.Classification technology is very popular in various automatic medical diagnostic tools.The diagnostic accuracy of some diseases has far exceeded that of human doctors.At present,the diagnosis and detection of liver diseases are mostly based on functional detection.In addition,it can analyze the level of enzymes in human blood for diagnosis.Through equipment,it can detect the physiological indicators of human body,such as alkaline phosphatase(ALP),total bilirubin(TB),direct bilirubin(DB),alanine transaminase(ALT)and aspartate transaminase(AST),albumin(AI)and total protein(TP)The ratio of.Traditional liver disease detection uses support vector machine(SVM)to establish classification model.The parameters of SVM are not selected to reach the optimal value of the model,which makes the testing time of the model too long and the accuracy of the model does not reach the expected effect.On this basis,the full intelligent algorithm is used to provide optimization support for the parameters of the vector machine,which makes the parameter optimization develop towards a more accurate direction.It can improve the accuracy of liver disease prediction model and shorten the detection time,so using swarm intelligence algorithm is an effective way to support the optimization of machine-oriented parameters.In this paper,the detection indexes of Indian liver disease patient data set provided by UCI machine learning database are taken as the research object,the features of data are extracted to build the model analysis,and the parameters performance of SVM is optimized through swarm intelligence algorithm,and the liver disease detection classifier is constructed based on these indexes and relations,so as to achieve better classification accuracy of SVM model It plays an important role in helping to predict and improve the accuracy of diagnosis of liver disease as early as possible.This paper is divided into two parts to study the application of swarm intelligence algorithm in the optimization of liver disease detection parametersFirstly,the principle mechanism and several improved forms of the related swarm intelligence optimization algorithm are described,and the systematic induction and summary are carried out to lay the foundation for their fusion and application.Secondly,the paper studies the limitation of the standard particle swarm optimization algorithm in solving the problem of multi peak,introduces the concept of dynamic grouping,and proposes the multi cluster particle swarm optimization algorithm.The topology of PSO is changed from static to dynamic,and the fitness function is tested with different parameters.In addition,the parameters of support vector machine are studied and analyzed,the parameters of group intelligent algorithm support vector machine are selected,and the optimization test of parameters is realized.Finally,the specific application scenarios are studied,and a hybrid parameter optimization method(GSA + dmgpso)is designed based on dmgpso algorithm and grid search method.The preprocessing of data set is completed,and the feature selection is realized by constructing ROC curve of single feature model of data attribute.The hybrid algorithm uses grid search algorithm to determine the approximate region of parameter solution,and then Combined with the support vector machine mesh search algorithm parameter optimization,dynamic grouping multi cluster particle swarm optimization algorithm in the disease detection model.Finally,the related work of parameter optimization of model support vector machine is completed.The test results show that the accuracy of classification can be significantly improved by providing support for the vector machine model.