Research On Data Analysis And Safety Mechanism In Smart Elderly Care

The traditional old-age care model can no longer meet the needs of today’s old-age care,and "smart old-age care" will follow,providing a new direction for old-age care services.At present,the security of elderly care data is still the focus of attention,and there are many loopholes in storage technology,such as easy tampering and easy stealing.Cardiovascular diseases have seriously threatened the health of the elderly.Therefore,the research on the security mechanism of elderly care data and the exploration and analysis of elderly diseases are particularly important.In the process of realizing smart elderly care,due to the particularity and huge amount of health data of the elderly,it is necessary to ensure the safety of the elderly health data before intelligent analysis.This topic is mainly divided into two parts for research,namely First realize the secure access and storage of elderly care data,and then use machine learning methods to intelligently analyze the data and predict diseases.In order to ensure the reliability and security of pension data,the alliance Blockchain structure is applied to data access.First of all,by adopting the PKI authentication system,users are authorized with different levels of access to data.Secondly,the data layer of the alliance chain,the NTRU algorithm is used to encrypt and authenticate its data.According to the calculation and derivation of the probability of the NTRU algorithm for decryption failure and brute force attacks under the maximum intensity,the optimal parameter settings suitable for the alliance chain scenario are determined.The simulation experiment results show that the security of the NTRU algorithm can be increased to the highest level given that decryption succeeds.Finally,figuring out the consensus bookkeeping node,through deploying smart contract to the alliance chain network,it builds up the PBFT algorithm-oriented alliance Blockchain model,and tests its performance.The experimental results show that when a data query request is made within a certain sending rate range,the overall delay of the system is low and can be limited within 40 ms.Through reprocessing and normalizing the open source cardiovascular disease data set provided by the Kaggle platform,and conducting univariate and multivariate exploratory analysis to the data,the analysis results show that the characteristics of systolic blood pressure,diastolic blood pressure,cholesterol and age are related to the existence of cardiovascular disease.After data analysis,four cardiovascular disease prediction models are established based on Logistic regression,GBDT,Random forest and Light GBM.The four prediction models are evaluated through the following three indexes: Precision,Recall,and F1-score,which shows that Light GBM prediction model has the best performance.Finally,the Light GBM disease prediction model is optimized by KNN.After repeatedly cross-validation,the results show that the performance of the model is improved through KNN optimization.