Using Federated Learning in Anomaly Detection and Analytics on Real-time Streaming Data of Healthcare

In order to reduce mortality in critical care units, it is essential to monitor a critically ill patient consistently in a hospital's intensive care unit. According to Harvard University research, medical errors cause nearly 5,000,000 deaths in India annually. The third most common cause of mortality, medical error is one of the top 10 killers. In the neonatal ICU, there are 0.7 medical errors per patient, compared to 1.5 errors per child in a pediatric emergency room. The prevalence of HF appears to be high in India, with estimates ranging from 1.3 million to 4.6 million, and an annual incidence of 491 600-1.8 million. Heart failure is a clinical illness that develops as the cardiac condition deteriorates. The prognosis and course of therapy for HF patients depend heavily on early diagnosis. Low staff-to-patient ratios, a lack of established standard medical protocol, and infrequent monitoring and evaluation of patient safety and quality indicators are the main causes of medical errors. Since human error is the main cause of medical errors, the appropriate technology may be used to reduce them. The goal of this project is to make use of the Federated Learning approach in health care to solve the problem of clinical deterioration and ultimately save lives.