Media Recommendation via Transcript-Social Post Text Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face difficulty in determining relevant and interesting media content suggestions based on limited information such as titles, creators, and genres of media content items they have already watched.
Innovation Solution
Generating a transcript of a media content item, receiving social network posts associated with other media content items, computing correlations between the transcript and post text, ranking posts, and presenting suggestions based on these correlations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If media content suggestions are based only on titles, creators, and genres, then the system is simple to operate, but the recommendation accuracy is insufficient
Solution Approach 1:
The patent introduces social network posts as an intermediary data source between the media content item and the recommendation system. These posts serve as a mediator that contains user reactions, comments, and discussions, providing rich contextual information that improves recommendation accuracy without requiring direct complex analysis of user behavior patterns.
Solution Approach 2:
The patent transitions from traditional recommendation dimensions (title, creator, genre) to a new dimension based on social network post text analysis. By incorporating textual data from social posts and computing correlations between transcript text and post text, the system adds a new analytical dimension that significantly improves recommendation precision.
2Measurement precision
If more data sources are incorporated for recommendations, then recommendation accuracy improves, but information processing time increases
Solution Approach 1:
The patent performs preliminary actions by pre-processing social network posts and computing correlations between transcript text and post text in advance. The system pre-analyzes the relationship between media content transcripts and social posts, storing these correlation data for rapid retrieval during actual recommendation generation, thereby reducing real-time processing requirements.
Data Source
AI summary
Methods, systems, and media for presenting suggestions of related media content are provided. In some embodiments, the method comprises: generating, using a hardware processor, a transcript of a first media content item; receiving one or more social network posts associated with one or more other media content items; computing one or more correlations between text in the one or more social network posts and the transcript; ranking the social network posts based at least in part on the correlations; and causing one or more suggestions to view the one or more other media content items associated with the one or more social network posts based at least in part on the rankings to be presented.