Identification method for call prediction of call center

The invention provides an identification method for call prediction of a call center. According to the method, a call problem of a call center is modeled into a graph model, nodes represent different parts of service of the call center, after a graph is serialized, a bidirectional long-short-term memory network is used for learning a time sequence mode of the service of the call center, and future call capacity is predicted. The method can effectively predict the future call capacity of the call center. Compared with a traditional time sequence prediction method, the method has better accuracy and stability. 本发明提出了呼叫中心呼叫预测的识别方法。该方法将呼叫中心呼叫问题建模为一个图模型，节点表示呼叫中心服务的不同部分，将图序列化之后，使用双向长短期记忆网络来学习呼叫中心服务的时间序列模式，并预测未来呼叫容量，结果表明，该方法能够有效地预测呼叫中心未来的呼叫容量。与传统的时间序列预测方法相比，该方法具有更好的准确性和稳定性。