计算机科学 ›› 2026, Vol. 53 ›› Issue (6): 19-29.doi: 10.11896/jsjkx.250600192
刘如意, 吕筱晗, 苗启广, 卢子祥, 王笛
LIU Ruyi, LYU Xiaohan, MIAO Qiguang, LU Zixiang, WANG Di
摘要： 在“互联网+教育”背景下,高校教育信息管理平台积累了海量学生行为数据,为学业预警研究提供了重要基础。然而,这些数据存在显著的类别不平衡问题。此外,由于学生行为数据以结构化形式存储,其特征间缺乏天然的空间关联性,传统深度学习方法难以有效挖掘特征间的潜在交互关系。此外,结构化数据中的不同特征具有明显语义差异,若不加区分地进行特征交互,可能导致模型学习到无效甚至误导性的关联模式,影响预警准确性。针对上述问题,提出了一种基于注意力机制和特征交互的学业预警预测模型。首先通过一种基于相对邻域密度的非线性少样本合成算法进行数据增强,以缓解类别不平衡问题。在模型架构上,采用残差连接结构,先通过可学习的多元高斯核将异质特征统一编码为向量表示,减少结构化数据中不同特征在数据类型和分布特性上的差异。在此基础上,结合语义匹配和注意力机制构建特征交互模块。具体而言,将每个特征建模为图中的一个节点,利用语义匹配机制动态计算边权重,并融合全局语义拓扑结构确定特征间的交互关系。此外,为直观展示特征对预测结果的影响,设计了基于泰勒公式的改进神经加性模块,利用泰勒多项式拟合输入输出间的非线性关系,使模型输出能显式呈现为特征的线性和非线性组合;同时引入张量分解技术,降低计算复杂度,提升高维数据处理效率,进一步提高了模型的泛化能力。
中图分类号:
| [1]ZHAO H N.Application of Data Mining in the Prediction of Student Retention in Universities [D].Lanzhou:Lanzhou University of Technology,2020. [2]ZHANG G P,QIN P Z,LI B W,et al.Research on Academic Early Warning Model in Higher Vocational Education in the “Internet Plus” Era:A Case Study of Jiangsu City Vocational College [J].Office Informatization,2024,29(3):24-26. [3]LUO J,XU S.NCART:Neural Classification and RegressionTree for tabular data[J].Pattern Recognition,2024,154:110578. [4]CHAWLA N V,BOWYER K W,HALL L O,et al.SMOTE:synthetic minority over-sampling technique[J].Journal of Artificial Intelligence Research,2002,16:321-357. [5]HAN H,WANG W Y,MAO B H.Borderline-SMOTE:a newover-sampling method in imbalanced data sets learning[C]//International Conference on Intelligent Computing.Berlin:Springer,2005:878-887. [6]HE H,BAI Y,GARCIA E A,et al.ADASYN:Adaptive syn-thetic sampling approach for imbalanced learning[C]//2008 IEEE International Joint Conference on Neural Networks(IEEE World Congress on Computational Intelligence).IEEE,2008:1322-1328. [7]BUNKHUMPORNPAT C,LURSINSAP C.Safe-level-smote:Safe-level-synthetic minority over-sampling technique for handling the class imbalanced problem[C]//Advances in Know-ledge Discovery and Data Mining:13th Pacific-Asia Conference. Berlin:Springer,2009:475-482. [8]FONSECA J,BACAO F.Geometric SMOTE for imbalanceddatasets with nominal and continuous features[J].Expert Systems with Applications,2023,234:121053. [9]LAVALLEY M P.Logistic regression[J].Circulation,2008,117(18):2395-2399. [10]YING L U.Decision tree methods:applications for classification and prediction[J].Shanghai Archives of Psychiatry,2015,27(2):130. [11]CHEN T,GUESTRIN C.Xgboost:A scalable tree boosting system[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2016:785-794. [12]BELGIU M,DRÃGUŢ L.Random forest in remote sensing:A review of applications and future directions[J].ISPRS Journal of Photogrammetry and Remote Sensing,2016,114:24-31. [13]KE G,MENG Q,FINLEY T,et al.Lightgbm:A highly efficient gradient boosting decision tree[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems.2017:3149-3157. [14]HE K,ZHANG X,REN S,et al.Deep residual learning forimage recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2016:770-778. [15]KLAMBAUER G,UNTERTHINER T,MAYR A,et al.Self-normalizing neural networks[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems.2017:972-981. [16]POPOV S,MOROZOV S,BABENKO A.Neural oblivious decision ensembles for deep learning on tabular data[J].arXiv:1909.06312,2019. [17]KATZIR L,ELIDAN G,EL-YANIV R.Net-dnf:Effective deep modeling of tabular data[C]//International Conference on Learning Representations.2020. [18]KE G,XU Z,ZHANG J,et al.DeepGBM:A deep learningframework distilled by GBDT for online prediction tasks[C]//Proceedings of the 25th ACM SIGKDD International Confe-rence on Knowledge Discovery & Data Mining.2019:384-394. [19]ARIK S Ö,PFISTER T.Tabnet:Attentive interpretable tabular learning[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2021:6679-6687. [20]HUANG X,KHETAN A,CVITKOVIC M,et al.Tabtrans-former:Tabular data modeling using contextual embeddings[J].arXiv:2012.06678,2020. [21]SUN L,ZHAO G,ZHENG Y,et al.Spectral-spatial feature tokenization transformer for hyperspectral image classification[J].IEEE Transactions on Geoscience and Remote Sensing,2022,60:1-14. [22]CHEN J,LIAO K,WAN Y,et al.Danets:Deep abstract networks for tabular data classification and regression[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2022:3930-3938. [23]HUI F G,RUI M T,YUN M Y,et al.Deepfm:A factorization-machine based neural network for CTR prediction[C]//IJCAI.2017. [24]VILLAIZÁN-VALLELADO M,SALVATORI M,CARRO B,et al.Graph neural network contextual embedding for deep learning on tabular data[J].Neural Networks,2024,173:106180. [25]CHEN J,LIAO K L,FANG Y,et al.Tabcaps:A capsule neural network for tabular data classification with bow routing[C]//The Eleventh International Conference on Learning Representations.2023. [26]WU J,CHEN S,ZHAO Q,et al.Switchtab:Switched autoen-coders are effective tabular learners[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2024:15924-15933. [27]AHAMED M A,CHENG Q.MambaTab:A plug-and-play mo-del for learning tabular data[C]//2024 IEEE 7th International Conference on Multimedia Information Processing and Retrieval(MIPR).IEEE,2024:369-375. |
|
||
首页