1

I am looking for recommendations for an open-source (clinical) decision support system that is based on ANN (artificial neural network, deep learning) technology. I would like to use it for my own education, but the license should also be permissive enough to allow possible commercial use at a later time. A Java-based implementation would be best. I am thinking of something like the output of an ongoing or past research project at a well-respected institution. I am mainly interested in the decision engine, much less so (except by way of example) in any domain-specific knowledge.

  • If I were building a clinical decision support system an ANN would not be the technology to use. A LCS (learning classifier system) would be my first stop. – Lyndon White Aug 28 '15 at 0:18
  • @Drux Did you ever end up finding one? – AdjunctProfessorFalcon Aug 8 '16 at 6:21
  • @Oxinabox Why is a LCS the better approach for a CDSS? – AdjunctProfessorFalcon Aug 8 '16 at 6:22
  • 1
    @Malvin9000 In short: A neural network is a tool for approximating functions, maps from one domain, to another domain, where both are expressed as numerical features -- this includes classifiers (A function from input features to 1 hot classes). It is very awkward to express all clinical situations and desired actions as tranformations from one numerical feature vector to an output numerical feature vector. Where as a LCS is a tool for recognising situtations and generating rules on what to do (from training data), and then selecting actions based on those rules. Try asking on CV stack ex. – Lyndon White Aug 8 '16 at 7:42
1

I would suggest taking a look at SciKit learn - it is Python rather than Java but has a very permissive, (BSD), licence and is actively used, developed and supported. It is fully open source and free.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.