The Design And Implementation Of Vertical Search Engine For Position Query

In recent years, more and more people have difficulties in applying for a job with the popularity of higher education. Consequently, some Internet companies find business opportunities and have established distinctive recruitment websites in succession, having built the bridge of communication between job seekers and recruiters. However, new problems have arisen in front of job seekers. In order to find a job, everyone usually needs visiting several recruitment websites to fulfill some tasks, such as user registration, job screening, resume editing, resume delivering and so on. It hurts people’s enthusiasm for looking for work. To simplify the process of job hunting, this paper has designed and implemented vertical search engine for position query. Not only it can gather the latest job information from each large recruitment website, but it can also choose the query results of meeting query intention from massive data for people, with taking advantage of the characteristics of fast, accurate, good real-time performance of vertical search. In the process of implementing the system, this paper firstly makes use of web crawler to grab data from various recruitment websites and put data-parsed into disk document libraries. Secondly, to improve the retrieval speed of the system, inverted index has been built with extracted data, but thinking of memory space constraints, several temporary indexes needs building and will be merged into final inverted index file. Thirdly, according to characteristics of the Boolean Model and Vector Space Model, the system will create a Boolean Retrieval Model with relevancy rank function. Finally, a basic web user interface has been designed, which can provide the query service of positions, companies and full text for users by calling job search API. After being evaluated by users, the proposed system can meet the query demands of job seekers at some level. However, some functions still need improving in the future research. On the one hand, dynamic update of job information needs to be resolved. On the other hand, sort function of retrieval model remains to be further improved.