摘要: 目前,我國高校都已經建立了較為全面的貧困大學生資助體系,但是由于學生的貧困生申請信息偏于主觀、貧困指標難以量化等因素,使得貧困生認定工作仍然是高校資助決策中的難點問題。尋求一種客觀、高效的貧困生認定評估標準,成為高校資助工作研究的重要內容。該文采用數據挖掘的手段,從學生的校園一卡通消費數據入手,使用K-means聚類算法對數據進行分析。在此基礎上,該文建立了基于聚類結果的貧困生指數算法計算每個學生的貧困生指數,用于輔助高校資助決策工作。
關鍵詞:數據挖掘;聚類;貧困生分析;校園一卡通 ;消費數據
中圖分類號:TP393 文獻標識碼:A 文章編號:1009-3044(2014)20-4934-03
Analysis of Impoverished College Students Based on Campus Card Consumption Data
FEI Xiao-dan1, DONG Xin-ke2,ZHANG Hui2
(1.School of National Defense Science and Technology, South West University of Science and Technology, Mianyang 621010, China;2.Network Information Center, South West University of Science and Technology, Mianyang 621010, China)
Abstract:At present, most of the universities and colleges in China have established a comprehensive system for aiding impoverished students. However, two of the factors accounting for the fact that identifying poor students is still a difficult problem are that the poor students application information is somewhat subjective and that the degree of poverty is difficult to quantify. Seeking an objective and efficient evaluation criterion for identifying impoverished students is one of the most important research themes in college funding. In this paper, data mining tools such as the K-means clustering algorithm are used to analyze campus card consumption data. In addition, based on the clustering result, an impoverished students index algorithm for calculating each student’s poverty index is established, which assists in decision-making of college funding.
Key word: data mining; clustering; analysis of impoverished students; campus card ; consumption data
我國已逐步建立起“獎、貸、助、補、減、免”等多種形式有機結合的較為完善的高校貧困生資助政策體系[1]。但是現有的基于人工的貧困生認定工作難以甄別申請材料的真實性,在認定中存在一定的主觀因素,同時也不能量化學生的貧困情況,如何客觀、高效地認定貧困生依然是一件十分困難的工作。
校園信息化建設的飛速發展使得利用學生平時在校的消費信息分析學生的消費行為成為了可能[2]。該文以西南科技大學一卡通消費數據為基礎,采用開源的數據挖掘工具weka進行二次開發來分析學生的消費行為,同時提出了K-means聚類算法下的貧困指數計算方法來輔助高校中的貧困生認定工作。
1 貧困生認定與校園一卡通消費數據
干凈而合乎要求的數據是數據挖掘成功應用的基礎[3],如何從海量的校園一卡通數據中獲取準確真實反映學生消費行為的數據成為貧困生分析的關鍵之一。……