移动通信中网络个性化服务信息提取仿真

Personalized Service Information Extraction Simulation Mobile Communication Network

对移动通信中网络个性化服务信息进行准确提取,可提高移动通信信息服务的质量。进行网络个性化服务信息提取时,需要预测用户对服务项目的未知评分,并计算用户的时间行为信息获得网络用户兴趣相似度,完成网络个性化服务信息提取。但是传统方法通过入挖掘潜在的网络用户社会关系,对移动通信中网络个性化服务信息进行提取,不能精确计算网络用户兴趣相似度,存在信息提取不准确、效率低的问题。提出一种关于用户兴趣度的移动通信中网络个性化服务信息提取方法。上述方法首先将移动通信过程中社交网络与用户评分矩阵组成一个矩阵,融合随机梯度下降理论对该矩阵进行分解,预测用户对服务项目的未知评分,获取相似用户对同一服务项目兴趣度的预测值与真实值之间的误差,并将用户访问时间引入社交网络服务提取过程中,采用用户的时间行为信息获得网络用户兴趣相似度,计算社交网络用户近似度和时间衰弱项,依据计算的结果对网络个性化服务信息进行提取。仿真结果表明,所提方法可以有效提高移动通信中网络个性化服务信息提取精度,网络开销比较稳定。

In this paper, we proposed an information extraction of personalized network service in mobile communication based on the interest similarity of network user. Firstly, we built a matrix consisting of social networking services （SNS） and user rating in mobile communication process. Then we decomposed the matrix integrated with stochastic gradient descent theory and predicted the unknown grade of user on service items to obtain the error between true value and interestingness prediction value of similarity user on the same service item. User access time was also introduced into the extraction process of SNS service. The interest similarity of network user was obtained using time behavior information of user. Finally, we calculated the similarity of SNS user and time weak item and extracted the information of personalized network service according to the calculation result. The simulation results show that the method mentioned above can improve the information extraction precision. The network spending is stable.