【作者】 曹路舟；

【导师】 周爱武；

【作者基本信息】 安徽大学 ， 计算机应用技术， 2011， 硕士

【摘要】 随着高职类院校招生规模的迅速扩大,贫困生数量也随之急剧增加,贫困生问题已成为学校学生工作的重要内容之一。然而,传统的贫困生认定工作通过这些年的工作实践后,发现存在很多的不足之处,迫切需要一套科学规范、易于操作的方法来完善贫困生认定工作,使贫困生认定工作能够高效、有序、合理地进行。因此,对高职类院校贫困生认定的研究有着重要的意义。高职类院校通过多年的发展,已经累积了大量的数据,如此多的数据在给教育工作者提供便利的同时也带来了很多的困惑,如何正确寻求隐藏在数据背后的有价值的信息以及发现蕴含在海量数据背后的潜在的联系和规则呢?数据挖掘技术可以帮助我们解决这个问题。数据挖掘技术融合了多个学科的知识,它能够从海量的数据中发现我们事先并不知道但却对我们有价值的信息。本文首先介绍了数据挖掘的基本知识,包括数据挖掘的概念、分类、过程及其挖掘常用技术；其次介绍了数据挖掘前的数据预处理,生成经过集成和转换处理后的数据信息总表；再次提出了关联规则的相关算法,包括Apriori算法和FP-growth算法以及在此基础上改进型的算法,接着利用这些算法对经过预处理的数据进行数据挖掘,生成相应的关联规则,对几种常用的关联规则算法进行了比较。从一定程度上来说,改进型的算法在寻求频繁项目集上减少了时间,尤其对数据仓库中巨大数据量进行挖掘的时候,效果明显,但是不管采用论文中介绍的哪种关联规则算法进行的数据挖掘所生成的关联规则结果应该都是一样的；最后把生成的关联规则结果与学校贫困生资助系统中实际的贫困生的相关信息作比较,分析挖掘效率,并说明产生这种结果的原因以及指出在以后贫困生认定过程中所要加强关注的方面。数据挖掘的技术还有很多种,而且每种挖掘技术的挖掘效率也不一定完全相同,针对具体的问题如何去选择好的挖掘算法,提高挖掘效率是今后研究的一个重要方向；而且目前高职类院校在学生管理工作中使用数据挖掘技术的地方还不多也不够深入,所以对数据挖掘技术在高职类院校贫困生认定工作中的应用研究有着广阔的前景。更多还原

【Abstract】 With the rapid expanding of higher vocational colleges’recruitment, the number of impoverished students are sharply increasing. The issue of impoverished students now has become one of the most important things in college.However, after many years practice, a lot of disadvantages have been found in the traditional endorsement of impoverished students; thus a set of scientific easy-to-use methods is urgently needed to improve the endorsement which can be processed efficiently and reasonably. So the research work of endorsement of impoverished students in higher vocational colleges is so meaningful.After many years development the higher vocational colleges have accumulated a large sum of data which provides convenience with confusion to the educators. How to correctly research the valuable information hidden behinde the data and find out the potential relation and regulation contained in the mass data? Data mining technique can help us to solve this problem. Data mining technique combines multiple disciplines of knowledge,it can help up from the vast amount ot data but that we do not know in advance of our valuable information.This thesis starts with the basic knowledge of data mining, including its definition, classification, process and some commonly-used techniques. Then it presents the general table of data information which is preprocessed, processed and converted before data mining. Then it introduces some algorithm of association rules which includes Apriori algorithm, FP-growth mining algorithm and the improved ones based on them in order to generate and compare the association rules that are mined from the preprocessed data. To some extent, the improved algorithums reduces the time in searching the frequent itemset, especially obvious in mining the huge amount of data in data warehouse. But whatever association rule is adopted,only the same result is generated. Last but not the least, it compares the result generated by the association rule with the actrual impoverished students’information in the colleges’ support system to analyse the mining efficiency and explain the reason of the result. The thesis also points out the aspects to which should be payed close attention in Endorsement of Impoverished Students.There are also plenty of data mining techniques with the same mining efficiency. To address specific problem,choosing the right mining algorithm to improve the mining efficiency is one of the most important subjects in the future. What’s more, data mining technique in the higher vocational colleges’students’ management work still needs to be further explored at present, which makes it a promising future in the research work of data mining in higher vocational colleges’ endorsement of impoverished students.更多还原