Abstract
Most published articles always applied a certain model or arithmetic to only a certain dataset. Considering the avalanche of biological data created in the post-genomic age, this type of research shows many shortcomings and inefficient characteristics, because it is always have difficulties to apply the same model to different datasets. So we proposed a multifunctional ensemble classifier which combines several individual classifiers. Each of them was trained in different parameter system. The final outcomes were combined through a weighted voting system. This classifier was conducted on several strictly constructed biological datasets. Based on the testing result from three different types of biological dataset, this new predictor can deal with more sweeping range of biological data, and receives more efficient and robust results in comparison with other published methods tentatively.
Keywords: Feature extraction, Fuzzy KNN, Protein quaternary structure, Pseudo amino acid composition, Subcellular location.
Graphical Abstract