Streaming Content Filtering by Viewer Bias Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Collaborative filtering systems in streaming services often recommend content that contradicts users' political biases, leading to user dissatisfaction and potential subscription cancellation.
Innovation Solution
Implement a content filtering system that classifies users into liberal, neutral, or conservative groups based on watch history, and removes or flags content items that contradict their biases, using machine learning models and collaborative filtering to tailor recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If collaborative filtering system recommends content items based on user preference and historical watch behavior, then content recommendation accuracy is improved, but user satisfaction deteriorates when recommended content conflicts with user biases
Solution Approach 1:
The patent extracts and removes content items from recommendations when they conflict with user biases. The system identifies biased content based on user watch history and explicitly excludes such content from recommendations, even though it would otherwise be accurate based on collaborative filtering algorithms.
Solution Approach 2:
The patent converts the harmful effect of biased content into a beneficial feature by using user watch history to identify and exclude biased content. The system leverages the same historical data that collaborative filtering uses for recommendations, but inverts the logic to remove rather than include biased items, thereby improving user satisfaction while maintaining recommendation accuracy.
2Object-affected harmful factors
If system filters out channels with contrary tags to user classification, then user satisfaction is improved, but content diversity deteriorates
Solution Approach 1:
The patent applies local quality by differentiating the treatment of content based on its alignment with user biases. Content is either fully included (aligned with user bias) or fully excluded (contrary to user bias), with no intermediate treatment. This creates homogeneous quality within the recommendation set while maintaining diversity in the overall content library.
3Reliability
If system removes channels contrary to user classification, then subscription retention is improved, but information loss increases
Solution Approach 1:
The patent converts the loss of contrary content into a benefit by using the same historical watch data to identify what content to remove. The system leverages the information about user viewing patterns to make intelligent exclusions, turning potential information loss into a targeted filtering mechanism that improves subscription retention.
Data Source
AI summary
Described herein are methods, systems, and media for filtering content items in a streaming service. In an embodiment, an exemplary method include classifying a user of a streaming service into one of a plurality of groups based on watch behavior of the user in a first past period, each of the plurality of groups being associated with a tag; and recommending channels to the user based on watch behavior of the user in a second past period, each of the plurality of channels being associated with a tag. The method further includes removing one or more channels from the recommended channels, each removed channel being associated with a tag contrary to the tag associated with the group that the user has been classified into; and displaying one or more content items from each of remaining channels on a streaming device of the user. The disclosure also discloses embodiments of flagging channels that conflict with user classifications.