The main goal of this exhibit is to introduce machine learning principles to those with no programming experience. Freely available software packages such as WEKA now make advanced machine learning techniques accessible to the non-programmer clinician. If clinicians learn the basic principles of machine learning, they may be able to identify problems that can be solved with machine leaning techniques. This can lead to more fruitful collaboration between clinicians and the machine learning experts.
After viewing this exhibit, the clinician should have a basic understanding of what machine learning is and what types of problems can be assessed with machine learning methods. The clinician should also be able to experiment with and explore machine learning using the WEKA software, as well as understand how to evaluate the effectiveness of a machine learning system.
What is Machine Learning?
Why should I bother leaning about machine learning if I am not a physicist or engineer?
Learning about machine leaning with Weka and a real mammography data set:
Introduction to Weka
Data visualization using Weka
Machine learning experiments with Weka
Evaluating performace: ten fold cross validation and area under the ROC curve
Information gain evaluation: relative importance of data features
Links
Raju, R,
Data Analysis Using Machine Learning Tools: An Introduction for the Non-programmer. Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL.
http://archive.rsna.org/2014/14014799.html
Selected for RadioGraphics