Figure 2.

Principles of machine learning Machine learning is a form of supervised learning in which a computer system learns from given positive and negative instances to distinguish between cases belonging to the two classes. During training, positive and negative cases (black and white balls) are provided for the system, which leads to organization of the predictor (indicate by the arrangement of the black and white squares inside the predictor) such that it learns to separate the cases and thus can classify unknown cases (balls with question marks). Depending on the classifier, whether it yields in addition to the classification also a score for the prediction, the results can be called as discrete or probabilistic.

Vihinen BMC Genomics 2012 13(Suppl 4):S2   doi:10.1186/1471-2164-13-S4-S2