摘要：【中文】本发明涉及电力行业的数据分析技术，具体涉及电网客服中心的话务量预测方法、系统、计算机设备及存储介质，其方法包括：获取呼叫中心在建模时间窗口的原始话务量数据，并对其进行标准化处理；对标准化后的话务量数据重构，计算重构序列的均值和方差，将重构序列及其均值、方差作为LSTM神经网络输入，设置神经网络输出层维数；建立LSTM神经网络模型，设置初始隐含层数和隐含层节点数量，训练、优化模型；将训练好的LSTM神经网络模型用于预测话务量，根据话务量数据的标准化处理过程，将预测变量逆向还原后与实际值对比，得到预测结果。本发明构建模型对话务量进行预测，精度较高、适应性较强，能够有效实现呼叫中心人力资源的最优配置。 【EN】The invention relates to a data analysis technology in the power industry, in particular to a telephone traffic prediction method, a system, computer equipment and a storage medium of a power grid customer service center, wherein the method comprises the following steps: acquiring original telephone traffic data of a call center in a modeling time window, and carrying out standardization processing on the original telephone traffic data; reconstructing the standardized telephone traffic data, calculating the mean value and the variance of a reconstructed sequence, taking the reconstructed sequence and the mean value and the variance thereof as LSTM neural network input, and setting the dimension of a neural network output layer; establishing an LSTM neural network model, setting an initial hidden layer number and a hidden layer node number, and training and optimizing the model; and (3) using the trained LSTM neural network model to predict telephone traffic, and comparing the predicted variable with an actual value after reversely restoring the predicted variable according to the standardized processing process of the telephone traffic data to obtain a predicted result. The method disclosed by the invention has the advantages that the model is constructed to predict the telephone traffic, the precision is higher, the adaptability is stronger, and the optimal allocation of the human resources of the call center can be effectively realized.

摘要：【中文】本发明涉及电网客户分群识别领域，具体为电网客户等级划分方法、系统、计算机设备及存储介质，其方法包括：从各个业务模块梳理建模所需源数据；从各个业务模块的各个指标维度设计相应的一个或多个业务指标，形成指标体系；基于指标体系对源数据进行数据加工、数据处理，得到业务因子指标，最终汇总成业务指标宽表；数据探索性分析，对业务指标的相关性进行分析，对业务因子指标进行有效筛选，生成业务指标；构建熵值法模型，计算并修正业务指标的权重，计算电网客户等级划分的最终得分，并划分电网客户群体。本发明构建熵值法模型，对业务指标的权重进行有效修正，以实现电网客户群体的精准划分。 【EN】The invention relates to the field of power grid customer grouping identification, in particular to a power grid customer grade classification method, a system, computer equipment and a storage medium, wherein the method comprises the following steps: combing source data needed by modeling from each business module; designing one or more corresponding service indexes from each index dimension of each service module to form an index system; performing data processing and data processing on the source data based on an index system to obtain service factor indexes, and finally summarizing the service factor indexes into a service index wide table; data exploratory analysis, namely analyzing the correlation of the service indexes, and effectively screening the service factor indexes to generate service indexes; and constructing an entropy method model, calculating and correcting the weight of the service index, calculating the final score of the power grid customer grade division, and dividing the power grid customer groups. According to the method, an entropy method model is constructed, and the weight of the service index is effectively corrected, so that accurate division of a power grid customer group is realized.

摘要：【中文】本发明涉及基于数据关联关系的检索路径分析与可视化系统及方法。基于数据关联关系的检索路径分析与可视化系统包括：关系网络可视化100、中间服务层200及图谱引擎300，所述图谱引300提供有客户端，该中间服务层200可以通过该客户端与所述图谱引擎300通信连接。优选方案，所述中间服务层200包括：关系网络解析装置210、节点属性查询接口230、图谱API接口220。本发明相对于现有技术的优点在于：建立业务逻辑模型与数据结构的映射关系，并通过图谱形式进行结果展现。 【EN】The invention relates to a retrieval path analysis and visualization system and method based on data association relation. The retrieval path analysis and visualization system based on the data association relationship comprises: a relationship network visualization 100, an intermediate service layer 200, and a graph engine 300, the graph engine 300 being provided with a client through which the intermediate service layer 200 can communicatively connect with the graph engine 300. Preferably, the middle service layer 200 includes: a relationship network parsing device 210, a node attribute query interface 230, and a graph API interface 220. Compared with the prior art, the invention has the advantages that: and establishing a mapping relation between the service logic model and the data structure, and displaying the result in a map form.