The Application Of Improved GA And RBFNN In The Prediction Of The Secondary Structure Of Protein

With the approach of post-genome era, proteomics is becoming an important research domain in the life science. The prediction of the secondary structure of protein is a challenging problem at present in the research of proteomics.This paper introduces the radial basis function neural network (RBFNN) one of the promising artificial neural network to predict the secondary structure of protein after introducing some methods about predicting the secondary structure of protein. The algorithm of choosing centers such as k-means, k-nearest neighbors about radial basis function neural network needs to be designated the number of categories , that the number of clustering . But the number of centers severely affects the performance of the neural network. So I use the genetic algorithm to cast about for the best centers. The genetic algorithm chooses the centers by wholly searching, decides the center vectors, spurns the added points and improves the performance of generality. The centers, breadth, parameter of structure and weights all affect the radial basis function neural network, so they are optimized by the adaptive genetic algorithm at the at the same time.The simple genetic algorithm is incapable of converging at the best point in all probability, at the same time, the crossover and mutation probability are decided subjectively. So the improved adaptive genetic algorithm is introduced to at the base of adaptive genetic algorithm. It availably conquers the subjectivity and uncertain about the simple genetic algorithm and speeds up the convergence and avoids prematurity.The encoding of the sequence of protein adopts the profile encoding which is abound with the evolvement message.In the inputting process it introduces sliding window to the prediction of the secondary structure of the homology protein.This paper mostly is about introducing improved adaptive genetic algorithm to radial basis function neural network in the prediction of the secondary structure of protein and obtaining fine conclusion.