Media Object Recommendation via Tag Overlap and Interaction Parameters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing media object recommendation systems often fail to provide the most relevant recommendations, as they rely on search ranking algorithms that do not adequately account for user interactions, leading to unsatisfactory results and increased user effort in finding desired content.
Innovation Solution
A method and system that utilize user interaction parameters and overlapping tags to determine the similarity between media objects, allowing for improved media object recommendation by selecting candidates with higher correspondence parameters, which are calculated based on user interaction values and tag overlap, thereby enhancing the relevance and efficiency of content suggestions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If search ranking algorithms are used for media object recommendation, then the system can provide recommendations based on general relevance, but the recommendations do not adequately account for user interactions leading to lower precision
Solution Approach 1:
The patent transforms the recommendation system by changing key parameters: introducing user interaction parameters (view count, play rate, engagement rate) as weighting factors, and modifying the correspondence parameter calculation to incorporate both tag overlap and user interaction metrics. This parameter transformation enables the system to account for user behavior patterns while maintaining the tag-based matching framework.
Solution Approach 2:
The patent introduces an intermediary correspondence parameter that bridges tag-based matching and user interaction analysis. This correspondence parameter serves as a mediator that combines structural similarity (tag overlap) with behavioral relevance (user interactions), allowing the system to generate accurate recommendations without directly comparing all possible user behavior data pairs.
2Measurement precision
If users manually search for relevant media objects, then they can find desired content, but the number of interactions and time required increases
Solution Approach 1:
The patent performs preliminary action by pre-calculating and storing user interaction parameters for all media objects in the database. When a recommendation request arrives, the system immediately retrieves these pre-computed metrics and combines them with tag overlap analysis, eliminating the need for users to perform multiple manual searches or for the system to analyze user behavior in real-time during the recommendation process.
Solution Approach 2:
The patent implements feedback mechanisms by incorporating user interaction parameters (view count, play rate, engagement rate) into the recommendation algorithm. The system continuously learns from user behavior patterns and adjusts recommendations based on this feedback, creating a closed-loop system that improves recommendation accuracy over time while reducing the need for manual user searching.
3Productivity
If traditional recommendation methods are used, then the system can provide basic recommendations, but the recommendations are not accurate enough requiring more user interactions
Solution Approach 1:
The patent applies the composite materials principle by creating a composite recommendation score that combines multiple different factors: tag overlap (structural similarity) and user interaction parameters (behavioral similarity). This composite approach integrates diverse data sources and measurement types into a unified recommendation metric, achieving high accuracy while simplifying the user experience as automatic recommendations replace manual searching.
Data Source
AI summary
A method of and a system for selecting recommended media objects comprising: acquiring media objects, each respective media object having at least one respective user interaction parameter, acquiring at least one tag associated with each respective media object, receiving, a request for a media object recommendation, acquiring at least one tag and at least one user interaction parameter associated with the media object, determining potential recommended media object candidates based on at least one tag of the media object overlapping with tags associated with the respective media objects, determining for each potential recommended media object candidate, a number of overlapping tags with the media object, generating a respective correspondence parameter based on the number of overlapping tags and the respective user interaction parameter, selecting at least one recommended media object, the at least one recommended media object being associated with a predetermined value of the correspondence parameter.