The Neural Network Method of Classifications for DNA Sequences
LI Yin-shan · Jisuanji fangzhen · 2003
This paper presents a method applying artificial neural network to DNA clustering problem. First we use the probability statistics method to extract the characters from the artificial DNA sequences whose categories are known. Thu s we can get the character vectors of the DNA sequences and input them as sample s into BP neuron NN for learning. We employ the BP(back propagation) algorithm t o train NN by use of the Neural Network Toolbox in MATLAB software package. In t his paper, two three-story NN are created to input the extracted DNA character v ectors as samples into them. After the training, characters are extracted from t he 20 unclassified artificial sequence samples and 182 natural sequence samples to form the character vectors as input of the two NN for clustering. The result s shows: the clustering method presented in this paper can classify the DNA sequ ences in quite high accuracy and precision. It is quite feasible to apply the ar tificial neural network to DNA sequence clustering.