基于搜索引擎的查询推荐算法研究  

Research on Query Recommendation Algorithm Based on Search Engine

作　　者：王晓迪 WANG Xiao-di(Software Engineering Department,Beijing University of Technology,Beijing 100124,China)

机构地区：[1]北京工业大学信息学部软件学院,北京100124

出　　处：《软件导刊》2020年第10期76-79,共4页Software Guide

摘　　要：传统的基于搜索日志的查询推荐方法无法快速有效处理和存储海量日志信息,无法抓住用户兴趣特点。为此,融合网络爬虫、数据挖掘和自然语言处理技术等多种方法,在原有查询日志数据基础上进一步爬取和挖掘,基于腾讯AI意图分析和自然语言处理技术,提出一种新的推荐词生成方法。实验结果表明,该方法与单纯基于查询意图的推荐和单纯基于相似度计算与聚类的推荐相比,用户查询准确性提升3%,能更加高效准确地为用户提供快速检索服务,提升了搜索引擎的用户体验。In order to improve the traditional search log-based query recommendation method which can not deal with and store massive log information quickly and effectively,and can not grasp the characteristics of user’s interests,this paper combines several methods such as web crawler,data mining and natural language processing technology,further crawling and mining on the basis of the original query log data.Based on Tencent AI intention analysis and natural language processing technology,a new method of generating recommendation words is proposed.The experimental results show that this method is more accurate and efficient than the recommendation based on query intention and the recommendation based on similarity calculation and clustering.By the method proposed in this paper,the query accuracy of users is improved by 3%,which significantly improves the user experience of search engines.

关 键 词：搜索引擎 查询意图 数据挖掘 爬虫 查询推荐 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]