基于时间窗序列的电力营销服务行为效率优化技术研究

Research on Optimization Technology of Power Marketing Service Behavior Efficiency Based on Time Window Sequence

为了提升电力营销业务量,促进电力营销行业发展,文章研究了基于时间窗序列的电力营销服务行为效率优化技术。通过计算电力营销服务行为数据向量的字符串获取意义串,建立时间窗口信息特征,通过定义动量模型属性获取突发话题。经离散余弦变换后,利用反向传播神经网络算法建立目标函数,获取实际输出值与期望输出值的误差均方值。依据误差均方值最小原则建立目标函数的凹解集合,通过梯度下降方法计算极小值点,完成电力营销服务行为效率优化。实验结果表明:该技术的数据预处理符合度始终保持在0.98左右,数据预处理能力较高;在时间为60个月内,优化后的人均业务量保持在19个左右,且销售变现时间明显降低,优化能力较强。

In order to improve the business volume of power marketing and promote the development of power marketing industry,the optimization technology of power marketing service efficiency based on time window series is studied.The meaning string of power marketing service behavior data vector is obtained by calculating the string,the time window information features are established,and the sudden topic is obtained by defining the momentum model attributes.After the discrete cosine transform(DCT),the objective function is established by using the backpropagation neural network algorithm to obtain the error mean square between the actual output and the expected output.The concave solution set of the objective function is established according to the minimum error mean square value principle,and the minimum point is calculated by the gradient descent method to complete the optimization of power marketing service behavior efficiency.The experimental results show that the data preprocessing conformity of the technology is always maintained at about 0.98,and the data preprocessing ability is high.Within 60 months,the optimized per capita business volume is maintained at about 19,and the sales realization time is significantly reduced,which indicates a strong optimization ability.