A Software Reliability Prediction Model: Using Improved Long Short Term Memory Network

With the development of software reliability research and machine learning, many machine learning models have been used in software reliability prediction. A long short term memory network (LSTM) modeling approach for software reliability prediction is proposed. Profit from its particular data flow control structure, the model overcomes the vanishing and exploding sensitivity of simple recursive neural network for software reliability prediction. Proposed approach also combines with layer normalization and truncate back propagation. To some extent, these two methods promote the effect of the proposed model. Compared with the simple recursive neural network, numerical results show that our proposed approach has a better performance and robustness with respect to software reliability prediction.