Research And Application Of Classification Method Based On Extreme Learning Machine

In recent years,the number of elderly population has increased year by year,the trend of aging population is more and more obvious,and the problem of guardianship for the elderly is urgently needed to be solved.However,the old-age services,which are mainly in hospitals and nursing homes,occupy limited medical resources,and the cost of nursing is huge.Obviously,to improve the efficiency of home health monitoring and reasonable allocation of limited medical resources needs to rely on advanced internet of things technology.Based on the internet of things,the wearable device is used to collect the daily life data of the elderly,track and judge the changes of the body state,finding out the bad trend in time,and taking the corresponding treatment measures to form closed-loop feedback health service.This technique can effectively solve the increasingly prominent problem of health care for the elderly,which is of great theoretical and practical significance.In this paper,the extreme learning machine classification method is mainly used to identify human behavior data in UCI machine learning database and to diagnose diseases from medical data,the classification method is also used to classify other popular data sets.In addition,some redundant features or noise data are processed by using heuristic search strategy and random search strategy combined with encapsulation strategy respectively.The main research contents are as follows: firstly,the classification principle of the extreme learning machine is analyzed,and a hybrid feature selection method combined with the extreme learning machine is proposed to realize the accurate classification of the reduced data.The search strategy and initialization method of artificial bee colony algorithm are improved,and the improved colony algorithm is tested on the standard test function,the results show that the improved algorithm has obvious improvement in precision and convergence speed.The improved multi-objective artificial bee colony algorithm optimizes the parameters of the ELM classifier model and feature vector at the same time,realizing the effective classification of the data,the results show that the algorithm has a good generalization ability.