Massachusetts Institute of Technology researchers say they have developed an algorithm that can accurately determine whether a patient is suffering from emphysema or heart failure based on readings from a capnograph (a machine that measures the concentration of carbon dioxide in a patient’s exhalations).
The researchers first identified features of the capnographic signal that appeared to vary between populations. For example, the crests of the waves in healthy subjects' capnograms seemed to plateau at a maximum concentration, while those in sick patients did not. After identifying about a dozen features, the researchers developed a machine-learning algorithm that would look for patterns in the features that seemed to correlate with patients' ultimate diagnoses.
However, the algorithm is somewhat unconventional in that the training data was split into 50 subsets. Each subset consisted of a random selection of about 70 percent of the data, meaning there was significant overlap between subsets, but no two subsets were identical. The researchers then used those subsets to train 50 different classifiers. The algorithm's ultimate output was the result of a vote by the 50 classifiers.
During testing, the researchers found their algorithm for distinguishing healthy subjects from those with emphysema yielded an area under the curve of 0.98. The algorithm that distinguished emphysema patients from those with congestive heart failure scored 0.89.
From MIT News
View Full Article
Abstracts Copyright © 2014 Information Inc., Bethesda, Maryland, USA