Pattern recognition using neural networks with a binary partitioning approach
L. Rudasi, Stephen A. Zahorian · 2002
The authors introduce a binary partitioned approach to classification which is applied to talker identification using neural networks. It was shown experimentally that the time required to train a single network to perform N-way classification is nearly proportional to the exponential of N. In contrast, the binary partitioned approach requires training times on the order of N/sup 2/. Evidence also exists to suggest that the binary partitioned neural network approach requires less training data than the use of a single large network. The binary partitioning approach was used to develop a talker identifier system for the 47 male speakers belonging to the Northern dialect region of the TIMIT database. The system performs with 100% accuracy in a text-independent mode when trained with about 9 to 14 s of speech and tested with 8 s of speech.>