基于频繁模式树和深度学习的频繁项集挖掘算法  

Frequent Itemset Mining Algorithm Based on Frequent Pattern Tree and Deep Learning

作　　者：李洋[1] 李华 Li Yang;Li Hua(The Clinical Medicine Department,Changchun Medical College,Changchun,Jilin 130031,China;College of Computer Science and Technology,Jilin University,Changchun,Jilin 130012,China)

机构地区：[1]长春医学高等专科学校临床医学院,吉林长春130031 [2]吉林大学计算机科学与技术学院,吉林长春130012

出　　处：《黑龙江工业学院学报(综合版)》2025年第1期94-98,共5页Journal of Heilongjiang University of Technology(Comprehensive Edition)

基　　金：教育部基金项目“基于高校开源数据的教学场景大数据分析模型构建研究”(项目编号:2023IT267)。

摘　　要：随着数据量的急剧增长,从海量数据中挖掘有价值的信息变得尤为重要。频繁项集挖掘作为数据挖掘的一个关键领域,旨在识别数据集中频繁出现的项集,这些项集能够揭示数据间的内在联系,并为后续的高级分析提供基础。然而,传统的频繁项集挖掘算法在处理大规模数据集时面临准确性和效率的挑战。为了解决这些问题,本研究提出频繁模式树和深度学习的新型频繁项集挖掘算法。该算法首先利用深度置信网络提取数据的高级特征,然后基于这些特征构建频繁模式树,以高效挖掘频繁项集。实验结果表明,该算法在查全率和查准率方面均表现优异,查全率高达97.56%,查准率高达95.49%,显示出其在实际应用中的高准确性和广泛适用性。With the rapid growth in data volume,it has become particularly important to mine valuable information from massive datasets.Frequent itemset mining,as a key area in data mining,aims to identify itemsets that occur frequently in datasets,which can reveal intrinsic relationships among the data and provide a foundation for subsequent advanced analysis.However,traditional frequent itemset mining algorithms face challenges in accuracy and efficiency when dealing with large-scale datasets.To address these issues,this research proposes a frequent pattern tree and a novel frequent itemset mining algorithm based on deep learning.The algorithm firstly utilizes deep belief network to extract high-level features from the data,then constructs frequent pattern trees based on these features to efficiently mine frequent itemsets.Experimental results show that the algorithm performs excellently in recall and precision,with a recall rate of up to 97.56%and a precision rate of up to 95.49%,demonstrating its high accuracy and wide applicability in practical applications.

关 键 词：频繁模式树 深度学习 频繁项集 数据挖掘 挖掘算法 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]