Video Streaming Quality Optimization Using Exit Rate Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video streaming quality evaluation and optimization methods fail to accurately reflect user experience, relying on metrics like initial frame delay, buffering duration, and frequency, which do not correlate well with actual user satisfaction, and lack real-time user feedback.
Innovation Solution
A method and system that utilize user behavior metrics, specifically exit rate and stay duration, to evaluate and optimize video streaming quality by calculating state transition probabilities and stay duration expectations, using an ϵ-greedy bandit algorithm to select optimal bitrates and CDNs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional metrics (initial frame delay, buffering duration, frequency) are used to evaluate video streaming quality, then the evaluation process is simple and objective, but the metrics do not accurately reflect user's real experience and viewing satisfaction
Solution Approach 1:
The patent applies feedback by using user behavior metrics (exit rate, stay duration) that directly reflect user experience to guide the optimization of video streaming quality. The system continuously monitors user behavior and uses this feedback to adjust bitrate and CDN selection, creating a closed-loop optimization system that accurately reflects user satisfaction while maintaining manageable complexity through automated data collection and analysis.
Solution Approach 2:
The patent changes the evaluation parameters from traditional technical metrics (initial frame delay, buffering duration) to user behavior metrics (exit rate, stay duration). This parameter transformation enables more accurate measurement of user experience while the system automatically handles the complexity of data collection and analysis, resolving the contradiction between measurement precision and system complexity.
2Productivity
If video bitrate and CDN are adjusted to reduce initial frame delay and buffering duration, then network transmission efficiency improves, but user viewing experience may not necessarily improve
Solution Approach 1:
The system uses user behavior metrics (exit rate, stay duration) as feedback to evaluate the actual impact of bitrate and CDN adjustments on user viewing experience. This feedback mechanism ensures that optimization decisions are based on real user experience rather than just technical transmission metrics, resolving the contradiction between productivity improvement and evaluation accuracy.
Solution Approach 2:
The patent replaces traditional mechanical measurement systems (technical metrics like initial frame delay) with a user behavior-based evaluation system. This substitution allows the system to measure actual user experience quality rather than just transmission efficiency, enabling accurate evaluation of whether bitrate and CDN adjustments truly improve user viewing experience.
3Measurement precision
If subjective surveys are conducted to obtain users' actual quality of experience, then accurate user experience data can be collected, but the process is time-consuming and costly
Solution Approach 1:
The patent applies self-service by using automatically collected user behavior data (exit rate, stay duration) that naturally occurs during video streaming. Instead of requiring active participation in subjective surveys, the system passively collects objective behavior metrics that automatically reflect user experience, eliminating the time and cost overhead of survey coordination while maintaining measurement accuracy.
Solution Approach 2:
The patent creates a copy of user experience through automated behavior metric collection. Rather than directly measuring subjective satisfaction through surveys, the system collects objective behavioral data (exit rate, stay duration) that serves as a proxy or copy of user experience. This copying approach maintains measurement accuracy while avoiding the time-consuming survey process.
4Measurement precision
If more comprehensive metrics and models are introduced to accurately evaluate user experience, then evaluation accuracy improves, but the system complexity and computational requirements increase
Solution Approach 1:
The patent extracts the essential elements of user experience evaluation into two key metrics: exit rate and stay duration. By taking out only the most critical behavioral indicators rather than incorporating all possible metrics, the system achieves accurate quality evaluation while maintaining relatively simple model complexity and computational requirements.
Solution Approach 2:
The patent segments the complex user experience evaluation problem into distinct components: exit rate measurement, stay duration measurement, and their combined use in quality assessment. This segmentation allows each component to be measured and processed independently, reducing overall system complexity while maintaining comprehensive evaluation accuracy.
Data Source
AI summary
Provided in the present invention are a video transmission quality evaluation and optimization method and system based on user behavior. The method comprises: establishing an exit rate model according to a session duration of a user watching a video, and a playing state of each second; calculating a state transition probability according to playing states of every two adjacent seconds; according to the transition probability and the exit rate, calculating mathematical expectation values of stay durations corresponding to positions, network connection types, CDNs and bitrates within different time periods; and according to a stay duration expectation, for the position and a network connection type of a given user, selecting a bitrate and a CDN which allow the longest stay duration expectation. By means of the present invention, a model is established on the basis of two user behavior metrics, i.e. an exit rate and a stay duration, and video transmission quality scores corresponding to different bitrates and CDNs are evaluated for the specific positions and network connection types of different users, such that a bitrate and a CDN which can optimize the viewing experience of a user are accurately given, and the video transmission quality can be greatly improved.