目的探索神经网络算法对数据缺失机制进行识别的可能性,为该领域的后续研究提供一定理论依据。方法通过建立神经网络模型,进行神经网络的学习与训练,来对数据进行分类识别,探索变量间可能存在的关系,进而进行缺失机制的识别。结果通过模拟研究与实证研究证实,神经网络模型可正确识别MAR缺失机制。结论神经网络模型应用于数据缺失机制识别具有一定的可行性,对于非MAR缺失机制中具体缺失机制的识别方法仍需进一步探索与研究。 Objective To explore the possibility of using neural network algorithm to identify data missing mechanism and provide some theoretical basis for the follow-up research in this field. Methods By establishing a neural network model, neural network learning and training to classify the data to identify possible relationships between variables, and then identify the missing mechanism. The results of simulation studies and empirical studies confirm that the neural network model can correctly identify the MAR deletion mechanism. Conclusion It is feasible to apply neural network model to data loss mechanism recognition. The identification of specific missing mechanism in non-MAR deletion mechanism still needs to be further explored and studied.