信息抽取技术在LBS中的应用  

Application of Information Extraction Technique in LBS

作　　者：张清军[1] 朱才连[1] 侯林山[1] 

机构地区：[1]中国科学院测量与地球物理研究所,湖北武汉430077

出　　处：《四川大学学报（工程科学版）》2005年第1期116-120,共5页Journal of Sichuan University (Engineering Science Edition)

基　　金：国家自然科学基金资助项目(40274058)

摘　　要：由于LBS系统的终端设备处理能力较低,显示屏幕较小,再加上无线数据网络带宽不足,因此无法浏览整个Web网页。采用信息抽取技术可以将用户感兴趣的信息提取出来,再发送给用户终端,有效地解决上述问题,信息抽取技术将是LBS系统中的一项重要应用。提出了一种基于信息抽取的从HTML到WML的页面转换方法,首先标记少量的Web网页形成样本实例集,采用归纳算法生成信息抽取规则;其次应用抽取规则和模式匹配来处理结构和风格类似的Web页面;最后将抽取结果转换为WML页面。开发了原型系统,通过对实际数据源的抽取,验证了此方法的有效性。Because LBS terminal devices have hardware constraints such as slow processing ability, small screen, and low bandwidth of wireless networks, it is difficult for these devices to display the entire Web page. It can effectively solve the above-mentioned problem that extracting the interesting contents for user from Web pages and send them to terminal devices. Therefore information extraction technique is an important application in LBS system. A new approach of page transformation from HTML to WML based on information extraction is put forward. Firstly, a set of training examples are generated from some Web pages labeled with examples of the data to be extracted, then extraction rules are induced from these user-labeled training instances; secondly, extraction rules and pattern match can be used to extract information from other Web pages similar to the training examples in structure and style; lastly, the extracted contents are transformed into WML page. A prototype system is implemented to test a set of Web pages. The experimental results show that this new method is effective.

关 键 词：LBS 信息抽取 模式匹配 页面转换 

分 类 号：P209[天文地球—测绘科学与技术] TP391.1[自动化与计算机技术—计算机应用技术]