In this paper we develop a robust classification mechanism based on a connectionist model in order to learn and classify objects from arbitrary feature spaces. Thereby a joint approach of recurrent neural networks and spread spectrum symbol encoding is implemented in order to classify any kind of objects that can be represented by feature vectors. Our main contribution is to adapt the spread spectrum method from signal transmission technology to classification of feature vectors, which are encoded for neural processing by means of unique spreading sequences. The idea behind this data spreading approach is related to the field of error-correcting output coding, but is furthermore characterized by a despreading mechanism that results in high classification accuracy and robustness against noisy or incomplete data. We applied our technique to four publicly available classification benchmarks, which stem from three different domains namely biology, geography and medicine. In the case of the MUSK2 molecule dataset (biology), ten-fold cross-validation of our technique revealed a classification accuracy of 97.7% at maximum, which is about 7% better than any published algorithm. In presence of a noise level of up to 25.0%, still an accuracy of 75.9% was achieved.
展开▼
