Weighted User Action Profile for Video Recommendation Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video on demand systems fail to accurately recommend content due to their inability to consider various factors that influence user preferences, leading to irrelevant suggestions.
Innovation Solution
A method that detects and identifies user actions on digital devices, relates these actions to digital content, and uses a sequencer and timer to form aggregate user actions, associating weight values with them, thereby evolving a user profile for personalized content recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If simple viewing history analysis is used, then system complexity is reduced, but recommendation accuracy deteriorates
Solution Approach 1:
The patent segments user actions into distinct types (play, pause, rewind, fast-forward, purchase, etc.) and processes each type separately with appropriate weight values. This segmentation allows the system to maintain complexity while improving accuracy by treating different user actions as meaningful individual data points rather than a simple viewing history list.
Solution Approach 2:
The patent adds temporal and hierarchical dimensions to the analysis by sequencing user actions in time order and organizing them into weighted categories. Instead of simple flat viewing history, the system creates a multi-dimensional profile that incorporates when actions occurred, what type of actions they were, and their relative importance weights, thereby improving recommendation accuracy without excessive complexity.
2Measurement precision
If multiple user action factors are considered, then recommendation accuracy is improved, but device complexity increases
Solution Approach 1:
The patent changes parameters by assigning different weight values to different user action types (e.g., purchase actions weighted higher than casual viewing). This parameter adjustment allows the system to consider multiple factors for improved accuracy while managing complexity through standardized weighting schemes rather than complex individual analysis of each action.
Solution Approach 2:
The system dynamically updates user profiles as new actions are detected, continuously adapting the recommendation model. This dynamic approach allows the system to incorporate multiple user action factors over time, improving accuracy while maintaining manageable complexity through incremental updates rather than complete reanalysis.
3Reliability
If real-time user action processing is implemented, then recommendation relevance is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary processing by immediately detecting and categorizing user actions as they occur, assigning weight values and updating profile segments in real-time. This preliminary action ensures recommendation relevance is maintained while managing processing time through efficient incremental updates rather than batch processing.
Solution Approach 2:
The system automatically processes user actions without requiring manual intervention, detecting actions, assigning weights, and updating profiles autonomously. This self-service approach improves recommendation relevance through continuous real-time processing while minimizing the time overhead by eliminating manual processing steps.
Data Source
AI summary
A user profile including content preferences is evolved by detecting and identifying user actions on network connected digital devices; relating the user actions to digital content; operating a sequencer and a timer to form aggregate user actions from the user actions; operating a second associator in conjunction with the timer and an adjuster to associate weight values with the aggregate user actions and the user actions, forming weighted user actions; and operating a profile server responsive to the second associator to evolve the user profile according to the weighted user actions.