Prediction of Health Care Data Using Efficient Machine Learning Algorithms

Abstract

Every clinical decision relies on the  doctor's expertise and comprehension. This standard  procedure may, despite appearances, lead to errors, biases,  and increased costs that compromise the patients' Quality  of Service (QoS).There is a pressing need for adaptable  equipment for critical patient care in developing nations  like India. The majority of Indian hospitals are unable to  provide their patients with adequate medical care due to a  lack of suitable, simple, and expandable intelligent  systems.The development of a comprehensive system that  will enable hospitals to provide vital patients with a real time feedback system is the objective of this project.Using  IBM cloud computing as a service platform and machine  learning, we propose a standard architecture, language, and  classification scheme for analyzing vital patient health data  (PaaS).The development of a machine learning (ML)  method for predicting a patient's fitness is the primary goal  of this study.Our models and data are stored and managed  by IBM Watson Studio and IBM Cloud.The Base  Predictors for our ml models are Nave Bayes, Logistic  Regression, the KNeighbors Classifier, the Decision Tree  Classifier, the Random Forest Classifier, the Gradient  Boosting Classifier, and the MLP Classifier.The precision  of the model has been increased by employing the  ensemble learning bagging strategy.We use a variety of  machine learning algorithms for ensemble learning.The  Critical Patient Management System, or CPMS, is a  mobile application we developed that allows for real-time  data and record viewing.Data that is relevant to ML model  training and deployment can be fetched in real time from  IBM Cloud and made available through CPMS because of  the way the system is built.Doctors can use Ml tendencies  to predict a patient's health status.The CPMS will send an  SMS notification to the duty physician and nurse to  provide immediate care if the situation worsens as  anticipated.Hospitals might get a smart healthcare solution  if the mission, milliliter models, and mobile application are  combined.