MRMD2.0: A Python Tool for Machine Learning with Feature Ranking and Reduction

TLDR: The study aims to find a way to reduce the dimensionality of the dataset by developing friendly python tools together with the web server that help to rank features and find the optimized dimensionality.

Abstract: \n\nThe study aims to find a way to reduce the dimensionality of the dataset.\n\n\n\nDimensionality reduction is the key issue of the machine learning process. It does\nnot only improve the prediction performance but also could recommend the intrinsic features and\nhelp to explore the biological expression of the machine learning “black box”.\n\n\n\nA variety of feature selection algorithms are used to select data features to achieve\ndimensionality reduction.\n\n\n\nFirst, MRMD2.0 integrated 7 different popular feature ranking algorithms with\nPageRank strategy. Second, optimized dimensionality was detected with forward adding strategy.\n\n\n\nWe have achieved good results in our experiments.\n\n\n\nSeveral works have been tested with MRMD2.0. It showed well performance.\nOtherwise, it also can draw the performance curves according to the feature dimensionality. If\nusers want to sacrifice accuracy for fewer features, they can select the dimensionality from the\nperformance curves.\n\n\n\n We developed friendly python tools together with the web server. The users could upload\ntheir csv, arff or libsvm format files. Then the webserver would help to rank features and find the\noptimized dimensionality.\n