Abstract: Performance appraisal in business administration has a great impact on social and economic development, so a sound performance appraisal system should be established. Moreover, in the information age, scientific methods are needed to improve business management performance. Based on this, this study links artificial intelligence with convolutional neural networks, and builds a corresponding performance research model based on actual conditions. When building the model, this paper selects the data width of 8Bit and 32 data per line, and shifts storage 2 rows, and sets the read/write enable signal to be half of the clock signal. In addition, the image matrix of the input image subjected to nonlinear processing by the excitation function ReLU will exhibit sparsity. Finally, combined with the model and data constructed in this study, the model is validated and the relevant strategies for performance evaluation are obtained.
