Protein secondary structure prediction: Speeding up conjugate gradient neural network

Protein secondary structure prediction from its sequence of amino acids remains an important issue. There are several methods devised to handle this issue. The most sophisticated method that has been devised for protein secondary structure prediction is artificial neural network (ANN). ANN can be trained using different learning algorithm. One of these algorithms is conjugate gradient (CG) learning algorithm which is popular learning rule for performing supervised learning tasks. In this paper principal component analysis (PCA) is used to reduce the dimensions of the neural network´s input vector, which in turn speed up the convergence in CG learning algorithm and thus, reducing the computational overhead. The results show that the number of epochs needed when reducing the data dimensions is about 30% of the number of epochs needed before reducing the data dimensions and the accuracy of prediction is increased by 1.2%.