A parallel feature selection based on rough set theory for protein mass spectrometry data

This paper presents an efficient parallel algorithm of optimal feature selection to reduce dimensionality for protein mass spectrometry data. The algorithm divides data into some parts to calculate separately, and then the relative importance of features is used for the parallel computing of each part. At last, the master process computes the final decision table reduction based on the part reduction. Experimental results show that the algorithm is suitable for mass spectrometry data. It not only reduces the computational cost but also keeps the classification accuracy.