Application Research Of Improved Deep Extreme Learning Machine In Lung Cancer TCM Syndrome Classification

Objective: The aim is to quantify lung cancer medical record data through feature selection and Likert grading,construct a deep extreme learning machine model based on sparrow search algorithm optimization,classify and forecast syndrome types of lung cancer TCM medical record data,and offer scientific and efficient methods for TCM syndrome classification research.And aiming at the problems arising from the use of a single and basic algorithm in most of the current traditional Chinese medicine data mining research,a platform(TCM data strategy model analysis platform,TCMDP)that optimizes traditional Chinese medicine data mining methods through strategic model intelligence is designed and implemented.Methods: A total of 497 cases of lung cancer diagnosed from January 2015 to December 2021 were collected from the Affiliated Hospital of Jiangxi University of Traditional Chinese Medicine,and 412 cases were screened as research objects.Feature selection and importance ranking were employed to summarize the syndrome factors of various syndrome types,and the Likert grading method was used to quantify them.Build a deep extreme learning machine optimized based on the sparrow search algorithm,and train and test the model.Finally,the model built in this study is compared with other machine learning models according to three evaluation criteria.TCMDP integrates the following four data mining modules according to the idea of the strategy model.The module of statistical analysis can analyze quantity,types,the meridian distribution of four and five flavors,and the efficacy of the drugs;conversely,the association rule mining module can examine the drugs in the prescription.Relevance;the cluster analysis module can obtain drug combinations to obtain clustered new prescriptions,and explore the compatibility rules of lung cancer prescriptions through the analysis results;the syndrome type classification module takes the TCM symptoms and four diagnosis information of the electronic medical record as input,and classifies the relevant syndrome types As an output,a syndrome classification model is constructed.To sum up,the intelligent data mining platform of traditional Chinese medicine based on the strategy model has been implemented,and the platform is used to study the medication rules and syndrome classification of traditional Chinese medicine prescriptions for the clinical treatment of lung cancer.Results: The average classification accuracy of the Deep Extreme Learning Machine Based on Sparrow Search Algorithm Optimization(SSA-DELM)model established in this study is 88.44%,and the average accuracy of the support vector machine and Bayesian network were 83.39% and 84.53%,respectively.The recall rate and F1 value of the SSADELM model on the five syndrome types are mostly above 80%,which is also better than other traditional machine learning models.As an illustration of TCMDP,the WD-Get Rule algorithm in the association rule mining module requires a minimum of 0.038 seconds to run when considering lung cancer with phlegm and blood stasis syndrome.The CMC-DD algorithm in the cluster analysis module takes a little longer to analyze,but the accuracy rate is as high as 87%.The running time of PSO-ELM in the lung cancer syndrome classification analysis module is 88.98 seconds shorter than other models,and the average accuracy rate of the model is 88.44%,which has a certain clinical reference value.Conclusion: The research results show that the lung cancer medical record data quantified by feature selection combined with Likert grading method can show the characteristics of the data better than the data processed by 0-1,and improve the accuracy of the classification model.The new SSA-DELM Compared with other traditional machine learning classification models,the model has better representation learning ability and learning speed.This model furnishes not only a scientific and technical approach to the clinical treatment of lung cancer,but also a beneficial reference for the data and astute formation of TCM syndrome differentiation and treatment.TCMDP’s advanced algorithms are far more efficient and precise than their predecessors,and can accurately choose the most suitable from the best;thus,utilizing strategy mode to construct and implement an intelligent mining platform of traditional Chinese medicine data is essential for the study of such information.The main work of this paper is as follows:(1)Determine the syndrome factors of different syndrome types by feature selection on the high-dimensional TCM case data,and combine the Likert classification method to quantify the data reasonably,and determine the syndrome factors of the five syndrome types.TCM syndrome differentiation and treatment is conformed to,the magnitude of TCM symptoms is adequately expounded,and electronic medical records are structured and clinical data standardized more effectively.(2)Use the sparrow search algorithm to optimize the input weight and hidden layer bias of the DELM model to reduce the impact of unstable classification results and low accuracy caused by random initialization parameters.This study’s adoption of the new SSA-DELM model has been demonstrated to be more precise than traditional models such as support vector machines and Bayesian networks in a variety of syndrome types.(3)The new model of TCM syndrome classification is constructed,which combines three methods of feature selection,Likert grading method and heuristic algorithm to optimize the neural network model,effectively dealing with the high-dimensional and nonlinear problems existing in the process of TCM syndrome classification,which reflects the complex relationship between symptoms and symptoms,symptoms and syndrome types in TCM medical record data,which not only provides effective ideas for the study of lung cancer syndrome differentiation,but also provides useful information and intelligent development for TCM syndrome differentiation and treatment.(4)The TCMDP platform is developed according to the idea of the strategy model.Integrating five analysis modules-statistics,association rules,cluster analysis,classification analysis,and syndrome diagnosis-the platform furnishes clinical treatment.It also furnishes references for medication for a variety of diseases,and is a valuable reference for studying the laws of TCM syndrome differentiation and treatment.