engleski

Machine learning approach identifies Baldwin effect during peptide immunomodulation in optic neuritis and multiple sclerosis

Empirical observations showed that standard statistical approach is not an appropriate tool for the determination of prognostic parameters during peptide therapy of immune-mediated diseases. We applied the alternative analysis based on the non-lonear machine learning approach with C4.5 decision tree classifier. C4.5 decision tree has been tested on the model of peptid-M (Lupex) therapy in optic neuritis/multiple sclerosis. The Baldwin effect has been observed through B and T cell population switching and selection. The accuracy of the procedure with the respect to therapy was 92-100% for small samples. The model of non-linear prediction provides useful alternative to the standard statistical approach, enables the extraction of few relevant parameters or their mutual relationships and ensures accurate prediction of the therapeutic procedures.

machine learning; Baldwin effect; peptide; immunomodulation; optic neutitis; multiple sclerosis; decision tree

nije evidentirano