Analysis of Audience Interest and User Clustering Based on Program Tags

다국어 초록 (Multilingual Abstract)

This paper proposes an analysis method of user behavior to provide personalized program recommendation based on program tags in the field of broadcasting and television. Multidimensional Scaling Analysis is used to produce a quantitative description of viewing preferences. Hierarchical clustering is performed to determine the number of clusters, followed by K-means clustering to group the data according to audience interest in TV program tags. This divides the audience into groups with similar viewing preferences.

목차 (Table of Contents)

• Abstract
• 1. Introduction
• 2. Audience Interest based on Program Tags
• 2.1. Delayering Tagging System of Television Programs
• 2.2. Audience Interest of Program Tags
• 3. Analysis of Audience Interest
• 3.1. Personal Tag Cloud of Audience Interest
• 3.2. Audience Multidimensional Scaling based on AIT
• 4. Audience Crowd Clustering based on AIT
• 4.1. Audience Clustering
• 4.2. Hierarchical clustering of Audience Interest of TV Program Tags
• 4.3. K-Means Clustering of Audience Interest of TV Program Tag
• 4.4. Simulation and Performance Analysis
• 5. Conclusion
• References