Home // BIOTECHNO 2020, The Twelfth International Conference on Bioinformatics, Biocomputational Systems and Biotechnologies // View article

A Versatile Combination of Classifiers for Protein Function Prediction

Authors:
Haneen Altartouri
Tobias Glasmachers

Keywords: protein classification; feature selection; clustering.

Abstract:
Protein classification problems can be addressed with a wide range of machine learning methods. Top performance is achieved with a variety methods, and the best method depends on the data set under study. Therefore, a minimal requirement for a general proceeding is to consider multiple classifiers, and to tune their hyperparameters. Further highly task-specific performance gains can be achieved through additional measures like feature selection, which is particularly important for high-dimensional descriptors, or with separate classifiers for different clusters. In this paper, we design a versatile classifier with the aim to combine all of the above options, but with robust defaults and fallback options. We demonstrate systematic performance improvements across a wide range of protein prediction problems.

Pages: 8 to 15

Copyright: Copyright (c) IARIA, 2020

Publication date: September 27, 2020

Published in: conference

ISSN: 2308-4383

ISBN: 978-1-61208-792-4

Location: Lisbon, Portugal

Dates: from September 27, 2020 to October 1, 2020