College Employment Recommendation Based on Improved K-Means Clustering and SimRank Algorithm in College Employment Management

Abstract

This research aims to tackle the talent development problem in universities by creating a smart employment recommendation system for recent college graduates. By combining an enhanced K-means clustering algorithm and SimRank algorithm, a new clustering center selection method has been implemented that focuses on maximizing the minimum distance. The proposed model addresses challenges related to sparse data and inaccurate suggestions, resulting in quicker processing times (299.8ms and 419.2ms) compared to four other clustering algorithms (UCF, RW, HCR, and SVD). Our model showcases superior clustering precision with minimal errors. After thorough system testing, we have identified the optimal parameter configurations: 25 recommended companies, a system recommendation hit rate of 0.68, and a system recommendation ranking index of 5.9. These results make a significant contribution to the field, presenting a highly efficient clustering model designed for employment recommendation platforms aimed at college graduates. Additionally, the study offers valuable insights for future research endeavors focused on improving job prospects for university students.

Keywords

• K-means
• SimRank
• college employment system
• employment recommendation
• cluster center