Architecture for using Machine Learning for Prediction of Pneumonia disease through Clinical Symptoms

Pneumonia is the biggest cause of death in children below the age 5 years, as well as the main source of death worldwide but mostly in developing countries. The detection of pneumonia is usually done by highly-trained doctors looking at a chest X-Ray radiograph. This is a time-consuming process that frequently results in radiologists disagreeing. Clinical-aided techniques showed promise in terms of increasing diagnostic precision. There was no single clinical characteristic that was sufficient to definitively diagnose pneumonia. The novel point of test care for the treatment of bacterial pneumonia can help to reach a sustainable amount of correctness, but adding clinical characters in a decision tree would improve treatment result. We use clinical symptoms like compress and vanishing deep twist machine learning, supplementations, and also we use aggregate learning to construct a computational strategy for detecting pneumonia in children. Early diagnosis can decrease the death rate in children. The WHO criteria for diagnosing clinical pneumonia are applied for the precision.