Cluster-Based Content Recommendation System for TV Guide Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face difficulty in selecting TV programming due to the large number of channels and content titles, and content may be unavailable if not pre-scheduled for recording.
Innovation Solution
A system and method that generates recommendations by classifying content into clusters based on multiple dimensional attributes, creating a taxonomy table, and comparing viewed content history to provide personalized internal recommendations lists on user devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If all channels and content titles are displayed in the grid guide, then users can see the full range of available content, but the screen cannot simultaneously display all channels due to space limitations
Solution Approach 1:
The patent segments the large set of content into multiple clusters based on categories, genres, or themes. Instead of displaying all content titles in a single overwhelming grid, the system divides content into manageable groups and presents them as clustered recommendations, allowing users to navigate through organized segments rather than facing the entire content library at once.
Solution Approach 2:
The patent introduces a new dimension of content organization by adding cluster-based categorization alongside the traditional channel and time slot dimensions. This creates a multi-dimensional navigation system where users can access content through both the conventional grid guide and the new cluster-based recommendation system, effectively expanding the information architecture without requiring additional screen space.
2Reliability
If users manually preschedule content for recording, then recording accuracy is improved, but users may miss content that is broadcast without prior scheduling
Solution Approach 1:
The system performs preliminary actions by automatically analyzing user viewing patterns and preferences in advance, then proactively scheduling recordings for recommended content without requiring explicit user commands. The system prepares recording schedules based on predicted user interest, ensuring content is captured before it is broadcast, thereby eliminating the need for manual prescheduling while maintaining high recording accuracy.
Solution Approach 2:
The patent implements a feedback loop where the system continuously monitors user viewing behavior, analyzes preferences, and adjusts recording recommendations accordingly. This feedback mechanism enables the system to learn from user interactions and improve its predictive scheduling accuracy over time, ensuring that automatically scheduled recordings align with user preferences while reducing manual intervention.
3Loss of information
If the grid guide displays detailed programming information for all channels, then information completeness is improved, but users find it difficult to decide on program selections due to information overload
Solution Approach 1:
The patent extracts and highlights the most relevant content recommendations from the complete set of available programming information. Instead of presenting all detailed programming data equally, the system identifies and extracts key recommendations based on user preferences and viewing history, presenting a curated subset that guides user decision-making without overwhelming them with excessive information.
Solution Approach 2:
The system applies local quality by providing different levels of information detail in different contexts. The cluster-based recommendations present summarized, high-level content groupings that are easy to scan, while the underlying complete programming information remains available for users who need detailed specifications. This localized information strategy ensures that users receive appropriate detail levels based on their immediate needs.
Data Source
AI summary
A method and system for recommending content includes a user device having a memory storing a taxonomy table having content cluster identifiers therein. The user device receives an external recommendations list for the content cluster at the user device. The recommendations list has a plurality of content identifiers each having one content cluster identifier. A viewer tracking module generates a viewed content history for content relative to the content clusters identifiers that correspond to viewed content at the user device. A recommendation module generates an internal recommendations list by comparing the external recommendations list to the viewed content history at the user device. The internal recommendation list also presents recommendations capturing the distinct user tastes in a family viewing device. A display displays the internal recommendations list, with section headers of different granularity describing the nature of the recommended content at cluster, sub-genre and genre levels.