Social Data Message Filtering for Inbox Delivery Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional messaging systems struggle to effectively filter out undesired content such as spam and abusive messages while often incorrectly filtering out desired messages, leading to a reduced user experience in social networking services.
Innovation Solution
The system determines whether to deliver a message to a primary or secondary inbox based on social data, including degrees of separation, privacy settings, user influence scores, common interests, and message content analysis, using a confidence score to assess the appropriateness of the message for delivery to the primary inbox.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If conventional filtering approaches are used to block spam and abusive messages, then some undesired content is filtered out, but many desired messages are incorrectly filtered out as well
Solution Approach 1:
The system changes the parameters used for filtering by incorporating multiple new dimensions: social graph distance (degrees of separation), group membership overlap, event attendance coincidence, location proximity, and common interests. Instead of relying on simple keyword filtering, the system evaluates messages based on the relationship between sender and recipient across multiple parameters, allowing desired messages to pass while blocking spam more accurately
Solution Approach 2:
The system creates a composite filtering approach by combining multiple social data sources (social graph, groups, events, locations, interests) into a unified confidence score. This composite evaluation method is analogous to using composite materials - each social data source contributes a different property to the overall message assessment, creating a more robust and accurate filtering system than any single data source could provide alone
2Measurement precision
If simple filtering rules are applied to messages, then processing speed is maintained, but filtering accuracy deteriorates
Solution Approach 1:
The system segments the filtering process into distinct evaluation components: social graph analysis, group membership comparison, event attendance matching, location proximity assessment, and common interests identification. Each segment evaluates a specific aspect of the sender-recipient relationship and contributes to the overall confidence score, making the complex filtering process more manageable and interpretable
Solution Approach 2:
The system introduces a confidence score as an intermediary metric that aggregates results from multiple social data sources. This confidence score acts as a mediator between the complex social data analysis and the final filtering decision, simplifying the output while maintaining the complexity of the analysis for accurate results
Data Source
AI summary
Systems, methods, and non-transitory computer-readable media can receive a request to send a message from a sender to a recipient. It can be determined that the sender and the recipient have at least a specified number of degrees of separation within a social network. The systems, methods, and non-transitory computer readable media can acquire social data about at least one of a set of groups joined by the sender and the recipient, a set of events attended by the sender and the recipient, a set of locations visited by the sender and the recipient, or a set of common interests shared by the sender and the recipient. It can be determined whether to deliver the message to a primary inbox of the recipient based on the social data.