Detecting Blood Clots using Neural Networks

The purpose of this research is to determine a noninvasive user-friendly method that the general population can use to determine whether they may have a blood clot. One thing that has impeded detection of blood clots in the past is the threat of invasive surgery as well as the cost of the tests that are offered. Our goal is to create a software system that takes blood clot symptoms as user input and provides the type of blood clot disease they are suffering from so that the patient can go to the hospital for further diagnosis. Many times people feel pain and they do not know the cause nor do they see a doctor to determine the problem. By allowing people to check themselves first in the privacy of their own homes it is possible that more people will see doctors with better knowledge of their potential problem. To accomplish this goal, we used a neural network model to determine the disease based on symptom inputs of the patient. The ´nftool´ of MATLAB with appropriate training provides an accuracy of 99.99%