Machine Learning for Protein Subcellular Localization Prediction

7 Ensemble random projection for large-scale predictions

Typically, GO-based methods extract as many as thousands of GO terms to formulate GO vectors. The curse of dimensionality severely restricts the predictive power of GO-based multi-label classification systems. Besides, high-dimensional feature vectors may contain redundant or irrelevant information, causing the classification systems suffer from overfitting. To address this problem, this chapter presents a dimensionality reduction method that applies random projection (RP) to construct an ensemble of multi-label classifiers. After feature extraction, the GO vectors are then projected onto lower-dimensional spaces by random projection matrices whose elements conform to a distribution with ...