Pneumonia Detection Using Data Mining Techniques

Abstract

As we know people in today's world face a lot of health issues it becomes essential to prevent these diseases before they occur. In this paper we dealt with one of the harmful diseases known as pneumonia. A comparison between the traditional methods and data mining methods for detecting pneumonia was performed. From various implementations we found that algorithms based on neural networks gave the best accuracies. The Learning algorithms were to perform the best in all the data mining techniques. These algorithms also include some pre-trained algorithms on X-ray images(chest). Transfer knowledge methodology of the pre-trained models gave the highest accuracy of 98% and a two-network architecture which is a combination of AlexNet and GoogLeNet gave an accuracy of 99%.