摘要
用户需求描述存在部分关联信息缺失及需求转变趋势模糊的问题,为此提出新的聚类模型.利用爬虫技术获取用户在线评论,依据Jieba分词结果构建异质词性有向链路网络,并基于PageRank算法计算词汇重要度,以重要度排序结果筛选特征词汇集.通过改进重叠模块识别算法,增强子群边属性相似度计算,构建评论特征词汇的无向加权网络;通过计算Jaccard距离开展网络图属性层次聚类,确定用户购买决策信息.通过多路径节点的相似性计算结果更新网络链路,实现用户需求预测.以水壶为例,改进的模块度为0.69,链路预测准确率达到86%,表明所提出的聚类模型能够明晰聚类结果潜在关联信息,链路预测结果符合群体效应客观特征,聚类结果可以辅助设计师明晰用户群需求以开展针对性设计工作.
A new clustering model was proposed to resolve the problem of missing associated information and ambiguous demand change trend in user demand description.The crawler technology was used to obtain user online comments,and the heterogeneous participatory link network was constructed with Jieba word segmentation results.The vocabulary importance was calculated based on the PageRank algorithm,and the PageRank calculation results were used as a basis for filtering to establish a feature vocabulary set.By improving the overlapping community discriminated algorithm,the similarity calculation of subgroup edge attributes was enhanced,and an undirected weighted network of comment feature vocabulary was constructed.By calculating the Jaccard distance,the hierarchical clustering of the network diagram attribute was carried out to determine the user’s purchase decision information.Update the network link with the similarity calculation results of multi-path nodes to achieve the user demand forecasting.Taking a kettle as an example,the improved modularity was 0.69,and the link prediction accuracy rate was 86%.Results show that the proposed clustering model can clarify the potential correlation information of clustering results,and the link prediction results are in line with the objective characteristics of the group effect.The clustering results can assist designers to clarify the needs of user groups in order to target their designs.
作者
张书涛
杨志强
王世杰
刘世锋
张凡
周爱民
ZHANG Shu-tao;YANG Zhi-qiang;WANG Shi-jie;LIU Shi-feng;ZHANG Fan;ZHOU Ai-min(School of Design Art,Lanzhou University of Technology,Lanzhou 730050,China;School of Mechanical and Electrical Engineering,Lanzhou University of Technology,Lanzhou 730050,China)
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2023年第9期1843-1855,1875,共14页
Journal of Zhejiang University：Engineering Science
基金
国家自然科学基金资助项目(51705226,52165033)
甘肃省青年博士基金资助项目(2022QB-047)
甘肃省高等学校创新基金资助项目(2021A-020)。
关键词
用户决策
复杂网络
社区发现
链路预测
user decision-making
complex network
community discovery
link prediction
