Preference Tag Segmentation for Accurate User Interest Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in accurately estimating user preferences from viewing history, as they often misinterpret genre preferences due to partial content interest rather than overall program preference.
Innovation Solution
An information processing device that assigns first and second scores to broadcast content based on preference tags and combined/single tags, respectively, to provide more detailed user preference estimation, using a combination of preference tags, combined preference tags, and single preference tags.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a single preference tag is assigned to broadcast content, then the classification is simple and easy to handle, but the accuracy of user preference estimation is insufficient
Solution Approach 1:
The preference tag system is segmented into three distinct types: preference tags (broad categories), combined preference tags (intersections of multiple preferences), and single preference tags (content matching only one preference). This segmentation allows the system to capture user preferences at different levels of granularity, improving estimation accuracy while maintaining manageable complexity through structured classification.
2Measurement precision
If multiple preference tags are assigned to broadcast content, then the accuracy of user preference estimation is improved, but the complexity of the system increases
Solution Approach 1:
Different types of broadcast content are assigned different types of preference tags based on their characteristics. News and documentary programs are assigned preference tags reflecting their informational nature, while entertainment programs receive combined preference tags capturing multiple appeal factors. This local differentiation optimizes the tag system for each content type, improving accuracy without uniformly increasing system complexity.
Solution Approach 2:
The system adds a new dimension to preference tagging by introducing combined preference tags that represent intersections of multiple preferences (e.g., news + documentary). This dimensional expansion allows the system to capture complex user preferences that cannot be expressed by single tags, significantly improving estimation accuracy while organizing complexity through a structured multi-dimensional framework.
3Measurement precision
If detailed preference analysis is performed, then the accuracy of matching recommended content with user interests is improved, but the amount of data required increases
Solution Approach 1:
The system changes the parameters of preference representation by using three distinct tag types with different levels of specificity. Combined preference tags allow the system to capture complex preferences with fewer data points, as they represent intersections of multiple preferences in a single tag. This parameter transformation enables detailed preference analysis with reduced data requirements compared to traditional single-dimensional tagging systems.
Data Source
AI summary
An information processing device includes: an obtainer that obtains a viewing history; a first evaluator (evaluator) that assigns a first score related to a preference tag assigned to each item of broadcast content; a second evaluator (evaluator) that assigns a second score related to a combined preference tag or a single preference tag; a first preference information table that stores the first score; a second preference information table that stores the second score; and an outputter that outputs preference information that matches a preference of a user in a predetermined period and includes one or more preference tags among (i) a plurality of kinds of preference tags, (ii) a plurality of patterns of combined preference tags, and (iii) a plurality of patterns of single preference tags.