油田信息搜索引擎个性化排序方法研究

发布时间：2018-03-04 01:37

本文选题：个性化排序　切入点：用户兴趣模型　出处：《东北石油大学》2013年硕士论文　论文类型：学位论文

【摘要】：相对互联网搜索引擎，企业搜索引擎用户进行搜索的目的性更强，对搜索结果排序的期望值更高，对同一检索请求希望根据用户的职业背景、兴趣爱好等得到个性化的检索结果。因此，构建一种个性化的排序方法是当前油田领域信息搜索引擎最迫切的需求。本文研究了个性化服务及搜索引擎排序的相关理论和技术，对油田领域多类型检索结果排序需求进行分析，提出了基于用户兴趣模型的个性化综合排序方法。本文的主要研究内容如下： 1．面向个性化排序的油田信息搜索引擎架构的研究。从油田搜索引擎的工作原理入手，研究了搜索引擎个性化处理方面的技术，提出了面向用户兴趣进行个性化排序的油田信息搜索引擎的体系架构。 2．研究并建立了油田搜索引擎用户的用户兴趣模型。采用用户主动提交兴趣信息和系统自动提取用户搜索历史与浏览行为信息相结合的方法获取用户兴趣信息，并基于改进的TF-IDF算法计算关键词权重，建立了用户兴趣向量空间模型；考虑了用户对搜索结果反馈度的问题，阐述了基于用户浏览行为和内容动态更新模型的方法和流程。 3．提出了基于用户兴趣模型的查询优化方法。提出了根据兴趣模型中用户历史查询和浏览行为计算用户兴趣度，，并结合扩展用词在结果文档中权重因素合理选择扩展用词对初始查询进行优化的方法，使得用户查询更加全面和准确。 4．为油田信息搜索引擎设计了个性化综合排序方法。研究了通用搜索引擎排序算法，针对它们的缺点问题进行改进，设计了一种基于用户兴趣模型的排序算法；然后分析了油田企业领域多数据类型检索结果的排序需求，为油田信息搜索引擎提出了将数据表结果优先排序，其它文档结果根据用户兴趣模型计算相关度后再进行排序的个性化综合排序方法。最后综合以上研究，阐述了面向个性化排序的油田信息搜索系统的设计和实现，以此验证了本研究能有效地提高油田企业搜索引擎查准率及满足用户的个性化检索需求，具有很大的现实意义和实用价值。
[Abstract]:Compared with the Internet search engine, enterprise search engine users have a stronger purpose of searching, and have higher expectations of ranking search results. For the same search request, they hope that according to the user's professional background, Therefore, constructing a personalized sorting method is the most urgent need of the oil field information search engine. In this paper, the relevant theories and technologies of personalized service and search engine sorting are studied, and the requirements of multi-type search results ranking in oil field are analyzed, and a personalized comprehensive sorting method based on user interest model is proposed. The main contents of this paper are as follows:. 1. The research on the structure of the oilfield information search engine oriented to the individualized sort. From the working principle of the oil field search engine, this paper studies the technology of the personalized processing of the search engine. The architecture of oilfield information search engine for personalized ranking of user interests is proposed. 2. The user interest model of the oil field search engine user is studied and established. The user's interest information is obtained by the combination of user's active submission of interest information and automatic extraction of user's search history and browsing behavior information. Based on the improved TF-IDF algorithm to calculate the keyword weight, the user interest vector space model is established, and the problem of user feedback to search results is considered, and the method and flow of dynamic updating model based on user browsing behavior and content are expounded. 3. A query optimization method based on user interest model is proposed, and the user interest degree is calculated according to user history query and browsing behavior in interest model. Combining the weight factor of extended words in the result document, the method of optimizing the initial query is chosen reasonably, which makes the user query more comprehensive and accurate. 4. The personalized comprehensive sorting method is designed for oilfield information search engine. The general search engine sorting algorithm is studied, and a sort algorithm based on user interest model is designed to solve their shortcomings. Then it analyzes the sorting requirements of multiple data types retrieval results in the field of oilfield enterprises, and puts forward the priority ranking of data table results for the oilfield information search engine. The personalized comprehensive sorting method based on the user interest model is used to calculate the relevance of other document results and then sort them. Finally, based on the above research, the design and implementation of the oilfield information search system oriented to individualized sorting are expounded, which verifies that the research can effectively improve the precision rate of the oil field enterprise search engine and meet the personalized retrieval needs of the users. It has great practical significance and practical value.
【学位授予单位】：东北石油大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.3

【参考文献】