TY - GEN
T1 - Ensemble Learning Based Employment Recommendation Under Interaction Sparsity for College Students
AU - Zhu, Haiping
AU - Zhao, Yifei
AU - Wu, Yuchen
AU - Chen, Yan
AU - Li, Wenhao
AU - Zheng, Qinghua
AU - Tian, Feng
N1 - Publisher Copyright:
© 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2023
Y1 - 2023
N2 - Recommendation systems play a crucial role in helping college students find job opportunities. However, the sparsity of interactions in employment recommendation for college students poses a challenge for models based on historical user preferences. To address this issue, we propose a novel model called Ensemble Learning based Employment Recommendation under Interaction Sparsity for College Students (EERIS). The model comprises two components: a similarity information component that uses pooled users to determine the nearest neighbor in user similarity measurement, and a global interaction component that uses interaction vectors of user groups to enhance interactions. To evaluate the missing interactions, we propose a loss function called CellLoss. These components are combined based on ensemble learning to improve the model’s generalization and scalability. Our experiments on two real-world datasets demonstrate the superior performance of the EERIS model. Ablation experiments further confirm that each component positively contributes to the model’s performance. Additionally, we design a revised metric for better model testing. Overall, the proposed EERIS model effectively addresses the interaction sparsity in employment recommendation for college students and provides satisfactory recommendations to students.
AB - Recommendation systems play a crucial role in helping college students find job opportunities. However, the sparsity of interactions in employment recommendation for college students poses a challenge for models based on historical user preferences. To address this issue, we propose a novel model called Ensemble Learning based Employment Recommendation under Interaction Sparsity for College Students (EERIS). The model comprises two components: a similarity information component that uses pooled users to determine the nearest neighbor in user similarity measurement, and a global interaction component that uses interaction vectors of user groups to enhance interactions. To evaluate the missing interactions, we propose a loss function called CellLoss. These components are combined based on ensemble learning to improve the model’s generalization and scalability. Our experiments on two real-world datasets demonstrate the superior performance of the EERIS model. Ablation experiments further confirm that each component positively contributes to the model’s performance. Additionally, we design a revised metric for better model testing. Overall, the proposed EERIS model effectively addresses the interaction sparsity in employment recommendation for college students and provides satisfactory recommendations to students.
KW - Employment recommendation
KW - Ensemble learning
KW - Interaction sparsity
KW - Recommendation system
UR - https://www.scopus.com/pages/publications/85177606580
U2 - 10.1007/978-3-031-46664-9_37
DO - 10.1007/978-3-031-46664-9_37
M3 - 会议稿件
AN - SCOPUS:85177606580
SN - 9783031466632
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 550
EP - 564
BT - Advanced Data Mining and Applications - 19th International Conference, ADMA 2023, Proceedings
A2 - Yang, Xiaochun
A2 - Wang, Bin
A2 - Suhartanto, Heru
A2 - Wang, Guoren
A2 - Jiang, Jing
A2 - Li, Bing
A2 - Zhu, Huaijie
A2 - Cui, Ningning
PB - Springer Science and Business Media Deutschland GmbH
T2 - 19th International Conference on Advanced Data Mining and Applications, ADMA 2023
Y2 - 21 August 2023 through 23 August 2023
ER -