Application of convolutional neural network in early detection of covid 19 through pneumonia.

The very rapid development of technology today cannot be separated from the role of artificial intelligence. Artificial intelligence is a thought process or intelligence like humans' that machines can use to do specific jobs. The use of artificial intelligence is very diverse in various fields, including the health sector. Recently the world's media has been shocked by the virus spread in China, namely the coronavirus. Corona can cause severe pneumonia for sufferers. Pneumonia itself is inflammation of the lungs caused by a virus. One of pneumonia sufferers' characteristics is that there is a fluid that looks like white spots on radiology or chest X-rays. However, it will be challenging to diagnose because at a glance, the radiological test results of lungs with pneumonia and normal lungs are similar. So, misdiagnosis is possible. Therefore, this study aims to create a machine learning model to classify lungs with pneumonia and normal lung on chest radiological images using the Convolutional Neural Network (CNN) algorithm. CNN is well known as a method in which the data process both image and video. The classification obtained for validation results in an accuracy value of 94%-99% for the pneumonia category, which suggests that the model can predict the image correctly. Meanwhile, for normal lung images, the accuracy rate is 84% -98%. This shows that the model can predict almost all of the images well; there is only one image that is not predicted correctly. [ABSTRACT FROM AUTHOR]