基于粗糙神经网络的用户轨迹数据深度挖掘

Deep Mining of User Trajectory Data Based on Rough Neural Network

利用目前方法对用户轨迹数据深度挖掘时,没有提取用户轨迹特征,存在终点预测精度低、系统模型鲁棒性低和系统模型命中率差的问题。于是提出基于粗糙神经网络的用户轨迹数据深度挖掘方法,通过对用户行为的影响因素进行分析与处理,获取用户轨迹数据特征,利用粗糙集简化数据获取挖掘规则;构建用户轨迹数据模型,采用聚类方法对规则数据产生的增长维数进行消减,最后确定模型输入,划分输入空间,为每个模糊划分选择合适的函数,完成用户轨迹数据深度挖掘。实验结果表明,通过对系统模型进行终点预测精度测试、鲁棒性对比测试和命中率对比测试,验证了该方法具有准确性高、可靠性高的特点。

During the deep mining of user trajectory data, the lack of user trajectory features leads to low end-point prediction accuracy, low robustness and poor hit rate of the system model. Therefore, we report a deep mining method of user trajectory data based on a rough neural network. The influencing factors of user behavior were analyzed and processed in order to obtain the characteristics of user trajectory data. A rough set was applied to simplify data and obtain mining rules. The user trajectory data model was established. The clustering method was introduced to reduce the growth dimension of rule data. Then, the model input was determined, the input space was divided, and the appropriate function was selected for each fuzzy partition. Finally, the deep mining of user trajectory data was achieved. The experimental results show that the method has high accuracy and reliability.