Dissertation > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer software > Program design,software engineering > Programming > Database theory and systems

Density-based Mining On Job Credibility

Author WangXin
Tutor HuangSui
School Jinan University
Course Applied Computer Technology
Keywords online recruitment data mining outlier detection DBSCAN LOF
CLC TP311.13
Type Master's thesis
Year 2013
Downloads 6
Quotes 0
Download Dissertation

With the rapid development of the Internet, online recruitment has become the mainmethod of human resources department. South China Market of Human Resources is themost professional regional online recruitment platform, which has been rapiddevelopment in the last ten years, but also encountered a series of problems, hindering thefurther development of it. Post information fidelity has been an important aspect to hinderthe development of South China Market of Human Resources, the problems also faced byother recruitment website. On the increasingly competitive environment of onlinerecruitment, who can improve the user experience, who will be able to occupy moremarket. As increase the credibility of the post is one important aspect.Outliers mining is an important topic of data mining. After ten years of development,it has become increasingly mature, and has been widely used in finance,telecommunications, as well as taxation and other industries. South China Market ofHuman Resources accumulated large amounts of raw data, including trusted and falsepositions after over a decade of practice. They identify the false post through users’complains and experienced experts, it not only takes a lot of manpower but also has acertain degree of subjectivity. Therefore, how to use outlier mining technology on thecredibility of the post is the focus of this study.To solve these problems, we studied the common methods of outlier detection, finallychose the two density-based outlier method to test(include DBSCAN method and the LOFmethod), and combine the two methods give a consolidated results. Finally, we use recalland precision to evaluate our experiment.

Related Dissertations
More Dissertations