Scholarlink

This framework presents a strategic model for video recommendation in the Web 3.0 era, integrating hybrid machine intelligence, generative AI, and semantic artificial intelligence through advanced semantic reasoning and quantitative semantic measures. The model dynamically generates ontologies from preprocessed query words, which are then used to select features based on linked similarity. These features are employed to classify the dataset from the perspective of the query using a logistic regression classifier. Semantics-oriented reasoning is achieved by computing the Normalized Pointwise Mutual Information (NPMI) to determine quantitative thresholds, and the Jian-Konrad Index is utilized as a criterion function for optimization. This optimization is carried out using the Elephant Optimization algorithm, which is executed only once to maintain the diversity and number of recommended entities. This approach ensures that the recommendations are both relevant and varied, aligning with the dynamic nature of Web 3.0.

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

VRSIL: A Framework for Video Recommendation Integrating Semantic Intelligence with a Large Language Model

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Anubrat Bora,
  • Gerard Deepak

摘要

This framework presents a strategic model for video recommendation in the Web 3.0 era, integrating hybrid machine intelligence, generative AI, and semantic artificial intelligence through advanced semantic reasoning and quantitative semantic measures. The model dynamically generates ontologies from preprocessed query words, which are then used to select features based on linked similarity. These features are employed to classify the dataset from the perspective of the query using a logistic regression classifier. Semantics-oriented reasoning is achieved by computing the Normalized Pointwise Mutual Information (NPMI) to determine quantitative thresholds, and the Jian-Konrad Index is utilized as a criterion function for optimization. This optimization is carried out using the Elephant Optimization algorithm, which is executed only once to maintain the diversity and number of recommended entities. This approach ensures that the recommendations are both relevant and varied, aligning with the dynamic nature of Web 3.0.