Content Recommendation via Common Class Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing content recommendation systems are limited in recommending diverse content options as they rely on previously reproduced content, restricting recommendations to specific categories.
Innovation Solution
An information processing apparatus and method that includes a content reproduction unit, a content part specification unit, a clustering unit, a class detection unit, a common class extraction unit, and a content retrieval unit, which classify and extract common classes from content parts to recommend content that aligns with a user's potential interests.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If content recommendation is based on previously reproduced content, then recommendation accuracy for specific categories is improved, but the diversity and range of recommended content is reduced
Solution Approach 1:
The content is segmented into multiple content parts (e.g., scenes, segments, or components) that can be independently classified into different clusters. This segmentation allows the system to extract multiple classes from the same content, enabling recommendations across diverse categories while maintaining accuracy for specific user preferences.
Solution Approach 2:
The patent introduces a new dimension of classification by extracting multiple classes from different clusters of content parts, rather than relying on a single category classification. This multi-dimensional approach expands the recommendation space while preserving precision in identified user preferences.
2Adaptability or versatility
If content is classified into multiple clusters, then the range of detectable classes is expanded, but the complexity of the classification process increases
Solution Approach 1:
The classification process is segmented into distinct stages: content segmentation into parts, clustering of content parts, and class detection from each cluster. This segmentation simplifies the overall complexity by breaking down the multi-cluster classification into manageable, sequential operations.
Solution Approach 2:
The patent extracts classes from each cluster independently rather than attempting to classify all content simultaneously. This extraction approach reduces classification complexity by processing each cluster separately and combining results, making the system more manageable while maintaining comprehensive class detection.
Data Source
AI summary
An information processing apparatus includes a content reproduction unit, a content part specification unit, a clustering unit, a class detection unit, a common class extraction unit, and a content retrieval unit. The content reproduction unit is configured to reproduce a reproduction content. The content part specification unit is configured to specify a plurality of content parts included in the reproduction content. The clustering unit is configured to classify the plurality of content parts into a plurality of clusters. The class detection unit is configured to detect a class from the plurality of content parts included in each of the plurality of clusters. The common class extraction unit is configured to extract a common class common to the plurality of clusters from the classes detected by the class detection unit. The content retrieval unit is configured to retrieve a content corresponding to the common class.