Scholarlink

<p>With the huge quantity of events published in event-based social networks (EBSN), it has turned out to be harder for the user to discover the events which well match their preference. Event Recommender System (ERS) emerges as a liable way to end this issue. Nevertheless, the ER scenario is rather dissimilar from the usual recommendation domain, since the events could not be “consumed” prior to their happening and sparse collaborative data. Even though certain works have emerged in this field, there is a lack of widespread study on the diverse features of EBSN data, which could affect ERS design. Thus, a novel technique is developed for ERS in social networks (SNs). Here, the input data are pre-processed using min–max normalization. After that, features like community willingness, personal willingness, improved Node Interest Degree, informative content feature and raw features are extracted. Further, improved PCA is deployed for the selection of noteworthy features. The step for ER in SNs is done with Deep Max out (DMO) and improved BI-GRU (I-BI-GRU) models. The outputs attained from DMO and I-Bi-GRU models are averaged to get recommended results on diverse events.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Hybrid model for event recommendation system with willingness, community and content feature set in social networks

  • 期刊论文

  • 发表日期：

    2025年10月13日

查看全文
  • Ashish Misra,
  • Vijay Singh,
  • Bhaskar Pant,
  • Arun Chauhan,
  • Ashwini Kumar Singh

摘要

With the huge quantity of events published in event-based social networks (EBSN), it has turned out to be harder for the user to discover the events which well match their preference. Event Recommender System (ERS) emerges as a liable way to end this issue. Nevertheless, the ER scenario is rather dissimilar from the usual recommendation domain, since the events could not be “consumed” prior to their happening and sparse collaborative data. Even though certain works have emerged in this field, there is a lack of widespread study on the diverse features of EBSN data, which could affect ERS design. Thus, a novel technique is developed for ERS in social networks (SNs). Here, the input data are pre-processed using min–max normalization. After that, features like community willingness, personal willingness, improved Node Interest Degree, informative content feature and raw features are extracted. Further, improved PCA is deployed for the selection of noteworthy features. The step for ER in SNs is done with Deep Max out (DMO) and improved BI-GRU (I-BI-GRU) models. The outputs attained from DMO and I-Bi-GRU models are averaged to get recommended results on diverse events.