Research on Anomaly Service Call Data Detecting Technology of Project Middle Platform Based on Bidirectional Long Short-Term Memory Network

The project middle platform has played an active role in the rapid construction of a new type of power system and the digital transformation of project management business at the State Grid Corporation. To achieve real-time control over the operational status of the project middle platform, it is necessary to carry out the detection of abnormal service call data. Based on the above issues, this paper proposes a detection method for abnormal service call data based on the Long Short-Term Memory (LSTM) network. Firstly, by analyzing the temporal characteristics of service call data, it is shown that the LSTM network can effectively extract its temporal patterns, and a double-layer fully connected LSTM network model is proposed. On this basis, combined with the weight allocation strategy of the attention mechanism, a detection method for abnormal service call data is proposed. Finally, the effectiveness and accuracy of this method are verified using real service call data, showing that the method has a good range of application. This paper has not yet considered how to correct abnormal service call data, and future research will be conducted in this direction.