基于BP神经网络的企业内部控制体系评价研究

对内部控制体系评价可以采用定性和定量两种评价标准。定性评价受评价人员的主观判断的影响，往往缺乏客观性，而定量评价则以其科学、精确、可比的特征受到审计机构的欢迎。本文构建了内部控制评价指标，把描述企业内部控制状况的特征信息作为神经网络的输入向量，而把代表相应综合评价结果的值作为神经网络的输出，并用足够的样本训练这个网络，使不同的输入向量得到不同的输出量值，从而对企业的内部控制状况进行了量化的评价。

The Research for Evaluating Enterprise Internal Control System on BP Neural Network

Both the nature and ration evaluation criteria car be used for evaluating internal control system. The nature appraisal is affected by personnel's subjective judgment, therefore,it lacks the objectivity. But the quantitative one is well received by the audit agency with its scientific, accurate and comparable characteristics. The article sets the index system of enterprise internal control, takes the charateristics that are used to describe the internal control of some interprises as input vector of the neural network and the comprehensive evaluating result as the output. Trained with enough specimens, the network will obtain different output values with different input vectors, thus appraising the state of enterprise control quantitativelg.