基于本体的苹果病虫害垂直搜索引擎研发

发布时间：2018-06-09 05:17

本文选题：苹果病虫害 + 领域本体　；参考：《西北农林科技大学》2017年硕士论文

【摘要】：随着互联网的迅速发展,网络信息的急剧膨胀给果农提供丰富病虫害信息资源的同时,又在有效使用方面面临新的挑战。针对现有搜索引擎查准率低、基于关键字的信息检索技术不能满足用户在语义上的需求,本文在领域本体的基础上展开对苹果病虫害垂直搜索引擎的相关研究。本文主要研究内容及结论如下:(1)苹果病虫害本体的构建。针对手工本体构建效率低,对领域专家依赖性太强等问题,采用基于文献的领域本体构建方法构建苹果病虫害本体。首先,本文对领域语料进行中文分词和数据清洗,通过相关性判断得到领域概念。然后,利用关联分析和聚类技术得到领域概念间的关联关系。最后,利用关系数据库对领域本体进行存储。该方法减少了对领域专家的依赖性,提高了领域本体构建的自动性。(2)苹果病虫害垂直搜索引擎的构建。针对苹果病虫害信息检索准确率低的问题,将领域本体引入垂直搜索引擎的构建中。首先,进行信息的采集和过滤。选取高质量的URL作为初始种子,用爬虫爬取相关页面并对页面进行分析,抽取出页面中的链接,并利用正则表达式对链接进行过滤,将过滤后的URL放入URL库;抽取网页正文进行主题相关性判断,将相关性高的网页存储在计算机中,按倒排索引策略动态建立索引。然后,实现信息的扩展查询和相关词的可视化推荐。最后对优化的主题爬虫、信息抽取性能和苹果病虫害垂直搜索引擎进行测试。实验表明,优化的爬虫速度提高了12.7倍。基于包装器的信息抽取的查准率保持在94.6%以上,F-指标性能在87.5%以上。本文研究的搜索引擎返回结果数目适当,相关度比较好,比普通搜索引擎高22%,比“农搜网”高5.2%。(3)原型系统的设计与实现。为了方便对信息检索平台的使用和管理,设计并实现了基于本体的苹果病虫害垂直搜索引擎原型系统,实现了领域本体的构建、网页的动态抓取、动态建立索引、扩展查询和相关词可视化推荐等功能。
[Abstract]:With the rapid development of the Internet, the rapid expansion of network information to fruit farmers to provide rich information resources, but also in the effective use of new challenges. In view of the low precision of the existing search engines and the fact that the keyword-based information retrieval technology can not meet the semantic needs of users, this paper studies the vertical search engine of apple diseases and insect pests on the basis of domain ontology. The main contents and conclusions of this paper are as follows: 1) the construction of apple pest and disease ontology. Aiming at the problems of low efficiency of manual ontology construction and too strong dependence on domain experts, a method of constructing apple pest and disease ontology based on literature was adopted. Firstly, the Chinese word segmentation and data cleaning of the domain corpus are carried out, and the domain concept is obtained by correlation judgment. Then, the correlation relationship between domain concepts is obtained by using association analysis and clustering techniques. Finally, domain ontology is stored by relational database. This method reduces the dependence on domain experts and improves the automatic construction of domain ontology. Aiming at the problem of low accuracy of apple pest information retrieval, domain ontology is introduced into vertical search engine. Firstly, the information is collected and filtered. The high quality URL is chosen as the initial seed, the crawler is used to crawl the relevant page and analyze the page, the link in the page is extracted, and the link is filtered by the regular expression, and the filtered URL is put into the URL library. The text of the web page is extracted to judge the relevance of the topic, and the pages with high correlation are stored in the computer, and the index is built dynamically according to the inverted index strategy. Then, the extended query of information and the visual recommendation of related words are implemented. Finally, the optimized subject crawler, information extraction performance and vertical search engine of apple pests and diseases were tested. The experimental results show that the optimized crawler speed is increased by 12.7 times. The precision rate of information extraction based on wrapper is above 94.6% and the performance of F- index is over 87.5%. The number of returned results of search engines studied in this paper is appropriate and the correlation degree is good. The design and implementation of the prototype system is 22% higher than that of ordinary search engines and 5.2% higher than that of "Nong Sou net". In order to facilitate the use and management of information retrieval platform, a prototype system of vertical search engine for apple diseases and insect pests based on ontology is designed and implemented. Extended query and related words visual recommendation and other functions.
【学位授予单位】：西北农林科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.3

【相似文献】