摘要
对电商服务过程中用户的需求进行准确预测,可以扩大电商服务范围,提高电商的服务质量。进行需求预测时,需要针对影响电商服务的不同因素,建立不同的预测指标,而传统的利用双重邻居选取策略,只能以单一指标为基础进行推荐行为的选取,降低了电商服务需求预测的准确性。提出一种本体的电商服务过程中用户需求均衡预测方法。首先构建电商服务均衡推荐用户本体,计算本体之间的语义相似度,得到相似用户集合,并将预测目标用户的近邻对象作为预测群,计算群中预测概率较高的信任子群,通过不确定近邻的动态度量进行平衡的预测,有效的完成了对电商服务中用户需求均衡预测。仿真结果证明,上述方法可以有效地提升电商服务中用户需求预测的准确性。
In this paper, an equalization prediction method of user requirement in the progress of electric commerce service is proposed based on ontology. Firstly, the suggested user ontology of electric commerce service was built to calculate semantic similarity between them. The similar user set was obtained. Then we took neighbor object of prediction target user as prediction group and calculated the trustworthy subset with high prediction probability in the group. Finally,we predicted the equalization through dynamic measurement of uncertain neighbor to complete equalization prediction of user requirement in electric commerce service. The simulation results show that the method mentioned above can improve precision of prediction effectively.
出处
《计算机仿真》
北大核心
2017年第2期426-429,共4页
Computer Simulation
关键词
基于本体
电商服务
均衡推荐
Based on ontology
Electrical commerce services
Balanced recommended