基于多约束推理的互联网地理位置信息挖掘算法研究

Internet Location Service Information Mining Algorithm Based on Multi-constraint Reasoning

现有的地理位置信息服务大都建立在地图测绘的基础上,随着移动互联网的迅速发展,利用互联网中包含的位置信息,成为了地理位置信息来源的一个新渠道。现有的地理位置信息挖掘算法主要是利用关键字匹配进行主题挖掘,然而在网络条件下的地理信息常常出现信息的错误和遗漏等情况,使得现有的基于关键字匹配的位置信息挖掘算法准确率迅速降低。针对这一问题,论文引入基于多约束的地理信息推理模型,将从网络中获取的多种标准化的信息作为约束条件进行推理,提升位置信息挖掘算法的准确性。实验表明论文提出的基于多约束推理的挖掘算法相比现有算法在地理信息准确度方面有明显提升。

Existing location information services are set up on the basis of mapping. With the rapid development of mo- bile Internet, the location information contained in the Internet, has become a new channel to the source of location informa- tion service. Current location information mining algorithm is mainly using keywords to the theme, but the geographic infor- mation under the network condition often appears information errors and omissions, making the accuracy of existing location information mining algorithm based on keywords matching reduce rapidly. Multi-constraint reasoning model has been intro- duced to solve this problem, it takes the standardization information obtained from the network as constraint conditions to improve the accuracy of location information mining algorithm. Experiment shows that the proposed mining algorithm based on constraint reasoning comparing with the existing algorithm has obvious improvement in accuracy of geographic informa- tion.