Prediction of Protein-protein Interaction Using Artificial Neural Network Back Propagation of Errors

This thesis explores the application of artificial neural network back propagation of errors to predicting protein-protein interaction. The back propagation-based method will produce a prediction result greater than 80% accurate rate. The high quality result will reduce the current high monetary cost, and time consumption of laboratory testing. In this paper, we will introduce our program called PPInteraction, and discuss the design and construction of an accurate dataset extraction method, including the training and testing of the artificial neural network back propagation application.