Random sequences of peptides in a microchip make it possible to generate specific immunosignatures that can diagnose various diseases. A large number of features does not allow for the quick and efficient analysis of such data. In this study, we propose technology to reduce feature space using various methods. The proposed technology makes it possible with minimal computational costs to ensure the accuracy and reliability of the classification of immunosignature data. The technology was tested on samples formed from a set of real data with the introduction of noise at various levels. The efficiency of the proposed technology on all test samples with various classifiers used for further data analysis is shown. .
Data repository URLs: