Enterprise invoice data monitoring method, device, computer device and storage medium

The present application relates to a method, apparatus, computer device and storage medium for monitoring enterprise invoice data based on machine learning. The method comprises: obtain real-time invoice data to be monitored, real-time invoice data is input into a pre-trained clustering model, When the offset of clustering center exceeds the threshold range, multiple historical invoice data whichare nearest to the real-time invoice data in the clustering model are obtained as samples of the nearest neighbor algorithm, and the monitoring results of the real-time invoice data are obtained according to the nearest neighbor algorithm and the identification tag to which the samples belong. The method can improve the accuracy of identifying abnormal types of invoices in invoice data. 本申请涉及种基于机器学习的企业发票数据监测方法、装置、计算机设备和存储介质。所述方法包括：获取待监测的实时发票数据，将实时发票数据输入预先训练的聚类模型，检测聚类模型的聚类中心的偏移量，当聚类中心的偏移量超过阈值范围时，获取聚类模型中与实时发票数据距离最近的多个历史发票数据作为最近邻算法的样本，根据最近邻算法以及样本所属的识别标签，得到实时发票数据的监测结果。采用本方法能够提高识别发票数据中异常发票的异常类型的准确性。