Aim: We aim to create a MATLAB implementation of the convex hull based feature selection algorithm, which is simple to use and easy to visualize data with. Users of the package will be able to train, test, and visualize data with customized option.
Social/scientific motivation: Due to rapid development of science and technology, the size and dimension of the data we need to analyse has become too large for human to understand. It has been necessary to introduce computational methods to analyse and interpret data. Thus, data mining and machine learning algorithms started to play an important role in discovering knowledge from these data.
Scientific background: The method is based on T. E. M. Nordling, N. Padhan, S. Nelander, and L. Claesson-Welsh (2015), “Identification of Biomarkers and Signatures in Protein Data” published in the Proc. of the IEEE 11th International Conference on eScience.
My motivation: It is a method with a simple concept, but it provides robust feature selection. It is a good way for me to learn about machine learning and statistics.
Members: Eric Lee