Digital Content Recommendation System Popularity Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital content recommendation systems often provide irrelevant content, known as 'clickbait,' which negatively affects user experience due to the difficulty in accurately assessing the relevance of non-native content items, as they may be geared towards attracting attention with sensational titles.
Innovation Solution
A method and system that compare system-specific popularity scores of non-native content items within the recommendation system with their web popularity scores to determine alignment, using a sigmoid function to calculate a probability value, and adjust ranking scores accordingly to filter out potentially irrelevant content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If non-native content items are included in recommendation to increase content diversity, then content variety is improved, but relevance accuracy deteriorates due to clickbait items
Solution Approach 1:
The patent introduces an intermediary verification mechanism that compares system-specific popularity scores with web popularity scores. This intermediary comparison acts as a mediator to identify clickbait items that have high system popularity but low web popularity, thereby filtering out irrelevant content while maintaining diverse recommendations
Solution Approach 2:
The patent implements a feedback loop where user interactions with recommended content are continuously monitored. The system uses feedback from user behavior patterns to adjust ranking scores and identify clickbait items, continuously improving relevance accuracy while maintaining content variety through iterative optimization
2Productivity
If ranking algorithms prioritize high-engagement content to improve user interaction, then user engagement is improved, but harmful clickbait content increases
Solution Approach 1:
The patent converts the harmful effect of clickbait into a beneficial filtering mechanism. By intentionally designing the ranking system to identify and flag high-engagement content that lacks corresponding web popularity, the system uses clickbait's inherent high engagement特性 as a signal for further verification, ultimately filtering out harmful content while maintaining legitimate high-engagement recommendations
Solution Approach 2:
The patent applies preliminary anti-action by proactively identifying and down-ranking potential clickbait items before they can significantly harm user experience. The system pre-calculates popularity alignment scores and establishes threshold criteria to prevent clickbait from rising to prominent positions in the recommendation list
3Device complexity
If popularity-based ranking is used to simplify content selection, then system complexity is reduced, but measurement precision of relevance deteriorates
Solution Approach 1:
The patent segments the popularity measurement into two distinct components: system-specific popularity (within the recommendation platform) and web popularity (external web presence). This segmentation allows the system to maintain simple individual measurements while achieving more precise relevance assessment through the comparison of both segments, effectively reducing overall system complexity while improving measurement precision
Data Source
AI summary
A method and a system for generating a digital content recommendation including a server configured to receive a request for the digital content recommendation. Based on the request, a set of candidate content items comprising a first content item and a second content item is generated. If the first content item is non-native to a recommendation system, the server is configured to determine if the first content item's popularity on the web and within the recommendation system align with each other. If there is a discrepancy above a threshold, a popularity adjustment score is assigned to the first content item. An adjusted set of candidate content item is generated based on the popularity adjustment score. The adjusted set of candidate content items is transmitted to the electronic device for displaying thereon.