Research on Software Reliability Evaluation Model Oriented Final Software Testing Process

In the traditional software reliability evaluation models, there are assumptions that the detected faults can be eliminated immediately and the new fault will not be introduced, which does not fit for the real final software testing process. This essay modifies the above assumptions, treats the detection and removing process of software as two separate process modeling, takes into account the influence of staging fault-removing and fault introduction rate, and then presents a software reliability model which is suitable for the real software testing and evaluation process. According to the actual situation, the appropriate the appropriate parameters are selected for the correlation function in the model. Two failure data sets are used to compare and validate the improved model, as a result, the improved model and the actual data have a good degree of consistency, fitting and forecasting effect is better.