Video Transcoding Priority Using Predicted Future View Numbers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video transcoding systems prioritize high-profile users, leading to delays in transcoding high-quality multimedia resources, resulting in suboptimal user experience due to the omission of important content.
Innovation Solution
Predict the future view number of multimedia resources based on submission features using a gradient boosting decision tree model to determine transcoding priority, ensuring timely transcoding of high-quality resources without relying on fixed parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the server prioritizes transcoding based on fixed parameters (e.g., high-profile users), then the transcoding process is simple to implement, but high-quality multimedia resources may be delayed or omitted, resulting in suboptimal user experience
Solution Approach 1:
The patent changes the parameters used for priority determination from fixed attributes (user profile levels) to dynamic predictive parameters (predicted view numbers based on submission features). This allows the system to identify high-quality resources more accurately by using multiple features including user history, resource type, and content characteristics, thereby resolving the contradiction between precision and complexity through more sophisticated parameter selection.
Solution Approach 2:
The patent replaces the mechanical rule-based priority system with a machine learning model (gradient boosting decision tree) that automatically learns optimal priority assignments from historical data. This substitution enables more accurate predictions of resource popularity while maintaining system automation, resolving the contradiction by using intelligent algorithms instead of simple fixed rules.
2Productivity
If the server processes all uploaded videos equally, then the system is easy to operate, but transcoding efficiency decreases and high-quality resources experience delays
Solution Approach 1:
The patent performs preliminary analysis of submission features and predicts view numbers before the actual transcoding process begins. This advance assessment allows the system to pre-determine priority levels and allocate transcoding resources efficiently, improving productivity by ensuring high-quality resources are processed first without requiring complex real-time adjustments during operation.
Solution Approach 2:
The system automatically evaluates submission features and assigns priorities without human intervention, using the gradient boosting model to self-determine which resources need urgent processing. This self-service mechanism maintains ease of operation while significantly improving transcoding efficiency through intelligent automated decision-making.
3Measurement precision
If the server uses a machine learning model to predict view numbers and determine priority, then transcoding specificity and precision improve, but the system complexity and computational requirements increase
Solution Approach 1:
The patent extracts only the essential features from submitted multimedia resources (user submission history, resource type, basic metadata) to feed into the gradient boosting model. By selecting and extracting only the most relevant features rather than processing all possible data, the system achieves high prediction accuracy while keeping the model structure manageable and avoiding unnecessary complexity.
Data Source
AI summary
Provided in the embodiments of the present disclosure are a multimedia resource processing method and apparatus, and an electronic device and a readable storage medium. On the basis of a contribution feature, the number of future visits of a multimedia resource uploaded by a contribution user can be predicted; a transcoding priority is obtained on the basis of the number of future visits of the multimedia resource; and the multimedia resource with a high transcoding priority is preferentially transcoded.