Video Playing Detection Method for Accurate User Engagement Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video playing behavior analysis methods are inaccurate due to issues like cyclic video playback inflating user video watching counts and inability to differentiate between video completion and skipping, leading to flawed analysis of user engagement and content quality.
Innovation Solution
A video playing detection method that determines video playback completion by analyzing the sequence of playing and quitting moments, introduces counting for quitting, jumping, and playback behaviors to accurately assess user engagement, and adjusts for abnormal cyclic playback.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If video playing counts of all users are simply summated at each moment, then the total count of playing the video by users at this moment can be obtained, but the accuracy of analysis on user watching behavior and favor degree is influenced due to cyclic playback inflation
Solution Approach 1:
The patent segments the video playing process into distinct states: playing state (count > 0) and non-playing state (count = 0). By tracking transitions between these states and recording quitting moments, the system can distinguish between actual viewing behavior and cyclic playback, thereby improving measurement precision while maintaining aggregation capability.
Solution Approach 2:
The patent performs preliminary recording of quitting moments and playing state transitions before final analysis. By pre-identifying when users quit viewing and when cyclic playback occurs, the system can adjust the aggregation counts to reflect true user engagement behavior rather than inflated cyclic playback numbers.
2Ease of manufacture
If only simple summation of playing counts is used, then the calculation process is simple, but the ability to differentiate between video completion and skipping is lost
Solution Approach 1:
The patent segments the playing count data into meaningful categories by tracking quitting moments and playing state transitions. This segmentation enables differentiation between users who complete videos, skip videos, or engage in cyclic playback, thereby preventing information loss while maintaining relatively simple calculation processes through state-based classification.
Solution Approach 2:
The patent introduces quitting moments as an intermediary element between raw playing count data and final behavior analysis. This intermediary recording mechanism preserves crucial information about when users stop viewing, enabling distinction between completion and skipping behaviors without requiring complex real-time analysis during playback.
3Quantity of substance
If playing count at a moment is zero, then no video playback is detected at this moment, but it cannot be judged whether user completes video watching or skips over video contents
Solution Approach 1:
The patent performs preliminary recording of quitting moments and playing state transitions even when current playing count is zero. By pre-tracking when users quit and maintaining state information, the system can later determine whether a zero count represents video completion or skipping, thereby preserving user behavior judgment capability without requiring non-zero playing counts for detection.
Solution Approach 2:
The patent introduces quitting moment records as an intermediary that bridges the gap between zero playing counts and meaningful behavior interpretation. This intermediary data structure enables the system to infer user behavior (completion vs. skipping) even when the immediate playing count is zero, preventing loss of behavioral information.
Data Source
AI summary
Disclosed in the disclosure are a video playing detection method and apparatus. The video playing detection method includes: obtaining a playing count of a video at a detection pending moment; obtaining a quitting moment of the video, wherein the quitting moment is a video playing completion moment; judging whether the playing count of the video at the detection pending moment is zero; if the playing count of the video at the detection pending moment is zero, judging whether the detection pending moment is before the quitting moment; and if the detection pending moment is before the quitting moment, determining that video playing is not completed at the detection pending moment. The disclosure solves the problem in the related art where a video playing behavior cannot be accurately analyzed.