Hello Everyone,

I have a simple yet a very complicated question. I am working in the filed of Healthinformatics, where I am processing a patients data that is coming from different sources like {blood test, urine test etc} and try to predict whether the person is healthy(0) or not healthy(1). So as you notices I have many inputs and binary output (0 or 1).

The question is which machine learning algorithms to use for my task apart from ANN I already using it? My thoughts are - Logistic regression - Support vector machines - Maybe fuzzy logic

Which one will better suite me and easier to use ?

Thank you

asked Mar 21 '14 at 23:58

SSU's gravatar image

SSU
1111

edited Mar 22 '14 at 00:11


One Answer:

I feel SVM suits you the best, considering that you will have to tune it the least and also it can give pretty go performance provided you dont have interdependencies etc in the input. You can use the grid search method on libsvm - binary classification.

answered Mar 22 '14 at 07:28

Sharath%20Chandra's gravatar image

Sharath Chandra
311131621

Your answer
toggle preview

powered by OSQA

User submitted content is under Creative Commons: Attribution - Share Alike; Other things copyright (C) 2010, MetaOptimize LLC.