应用模块化方法进行副学士录取候选人数据采掘的案例研究
Case study: Data Mining of Associate Degree Accepted Candidates by Modular Method
摘要:
大约10年前,伊朗应用科技大学(UAST)开始采用模块化方法招收非连续副学士学位学生,每年招收近10万名学生。虽然举办这些课程的首要目的是提高雇员的科学和技能水平,但随着时间的推移,有相当多的失业者对参加这些课程感兴趣。针对这一事实,本文采用无监督学习和监督学习范式,对2008和2009模块化课程录取考生样本数据进行了挖掘和分析。在第一步中,通过使用无监督范式,我们根据学生身份对模块化接受的候选人进行分组(聚类),并将数据集标记为三个类,以便每个类以某种方式显示模块化接受的候选人的教育和学生身份。在第二步中,通过使用监督和非监督算法,我们生成预测模型在2008年的数据集。然后,通过对生成模型的性能进行比较,选择了关联规则预测模型,并利用该模型提取了部分规则。最后,该模型被执行的测试集,其中包括接受的候选人的下一个课程,然后通过评估的结果,得到的结果的正确性和机密性的百分比可以查看。
英文摘要:
Since about 10 years ago, University of Applied Science and Technology (UAST) in Iran has admitted students in discontinuous associate degree by modular method, so that almost 100,000 students are accepted every year. Although the first aim of holding such courses was to improve scientific and skill level of employees, over time a considerable group of unemployed people have been interested to participate in these courses. According to this fact, in this paper, we mine and analyze a sample data of accepted candidates in modular 2008 and 2009 courses by using unsupervised and supervised learning paradigms. In the first step, by using unsupervised paradigm, we grouped (clustered) set of modular accepted candidates based on their student status and labeled data sets by three classes so that each class somehow shows educational and student status of modular accepted candidates. In the second step, by using supervised and unsupervised algorithms, we generated predicting models in 2008 data sets. Then, by making a comparison between performances of generated models, we selected predicting model of association rules through which some rules were extracted. Finally, this model is executed for Test set which includes accepted candidates of next course then by evaluation of results, the percentage of correctness and confidentiality of obtained results can be viewed.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。