IOT Based Clinical Data Set Formation Model for Smart Health Care Application

The rapid development of the Internet and computer technology makes it possible to develop IoT based health care services in today's ageing population. However, the IoT data problem in elderly care seriously restricts the process of intelligentization in this field. Issues such as missing real IoT data, interference of dirty IoT data and too few standard samples emerge in an endless stream. Aiming at the problem of missing IoT data sets, based on collecting a small sample IoT data of community health care in a city, a three-stage IoT data generation model based on machine learning is proposed. The generation strategy based on tree structure is used in the first stage. The distribution of the IoT data generates the essential attributes of the IoT data set; in the second stage, the naive Bayesian algorithm is used to generate the primary behavioural ability evaluation index of the sample; in the third stage, the multiple linear regression method is used to generate the high Behavioral competency indicators and evaluation stages. In addition, to verify the effectiveness of the IoT data set generated by the model for downstream tasks, based on the generated IoT data, this paper uses neural networks to design multiple rehabilitation training plan recommendation models to achieve five multi-classification tasks and two multi-label classification tasks. Finally, this paper verifies the authenticity and validity of the generated IoT data by analyzing the experimental results and injecting expert knowledge.