The prediction model of multiple myeloma based on the BP artificial neural network

The paper presents a prediction model of multiple myeloma by analysis of data sets of microarray of expression profile based on the theory of Back Propagation artificial neural network. This three-layer network model whose input is the expression status of 24 feature genes which picked out by statistical method and results of others’ study is self-learning and self-adaptive. The results of experiments show: with proper training the model can predict the samples of experimental data accurately.