| 摘 要: 随着高等教育的普及,各类管理决策需求增加,基于数据驱动和知识驱动的决策方法成为新的发展趋势。基于此,提出了一种通过改进概率软逻辑模型挖掘知识规律和推理学生状态的方法,该方法通过引入规则挖掘算法,赋予概率软逻辑模型自动挖掘规则的能力。在真实学生状态数据上的实验结果表明,该方法在多项目标关系的推理上,准确率能达到80%以上,并且多层次推理结构的实验效果优于非层次推理结构的实验效果。对于学生状态类非完整知识库数据的挖掘和推理,该方法减少了人工工作量,降低了因人为认知偏差带来的不确定性。 |
| 关键词: 概率软逻辑;高校学生状态;规则挖掘;推理;多层次结构 |
|
中图分类号: TP181
文献标识码: A
|
| 基金项目: 中国(绵阳)科技城网络应急管理研究中心项目(WLYJGL2023YB03[西南科大校内编号:23sd4117],WLYJGL2023YB01[西南科大校内编号:23sd4115]);西南科技大学学生教育管理及改革专项科研项目(23XGB016,23XGB019);西南科技大学素质类教改(青年发展研究)专项资助项目(23szjg11) |
|
| Inference Method of College Student Status Based on Improved Probabilistic Soft Logic Model |
|
ZHANG Jia1, WANG Zhixing2, WANG Jiao1
|
(1.School of Computer Science and Technology, Southwest University of Science and Technology, Mianyang 621010, China;
2.School of Civil Engineering and Architecture, Southwest University of Science and Technology, Mianyang 621010, China)
1005974319@qq.com; 602457064@qq.com; 171833221@qq.com
|
| Abstract: With the popularization of higher education, the demand for various management decisions has increased, and decision-making methods based on data-driven and knowledge-driven approaches have become a new trend. In this context, this paper proposes a method for mining knowledge patterns and reasoning student status through an improved probabilistic soft logic model. By introducing rule mining algorithms, this method endows the probabilistic soft logic model with the ability to automatically mine rules. Experimental results on real student status data show that this method can achieve an accuracy rate of over 80% in reasoning with multiple target relationships. Moreover, the experimental results of a multi-level reasoning structure are better than those of a non-hierarchical reasoning structure. For mining and reasoning incomplete knowledge base data of student status, the proposed method reduces manual workload and decreases uncertainty caused by human cognitive biases. |
| Keywords: probabilistic soft logic; college student status; rule mining; inference; multi-level structure |
