Click Fraud Detection via Delayed Rating Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing click fraud detection systems fail to prevent fraudulent clicks effectively, as they rely on identifying and blocking dangerous IP addresses after the fact, do not account for consumer intelligence, and cannot identify suspicious users until they have already committed fraud, leading to contaminated content and recommendation systems and financial losses.
Innovation Solution
A method and system that delay updating internal ratings until a user is determined trustworthy based on a user trust score calculated from various data points such as contract score, profile overlap, purchase history, and user behavior, and require approval from other trustworthy users before updating ratings, thereby preventing fraudulent click inflation and maintaining accurate content recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing click fraud solutions use IP address blocking to detect fraudulent clicks, then fraudulent activity can be identified after occurrence, but the system cannot prevent fraud before it affects ratings and content recommendations
Solution Approach 1:
The system performs preliminary actions by building user models and calculating trust scores before users can commit fraud. Trust scores are computed in advance based on user profiles, behavior patterns, and historical data, enabling the system to prevent fraudulent clicks before they contaminate ratings or recommendations.
Solution Approach 2:
The patent implements beforehand cushioning by maintaining a trust score buffer that protects the rating system from fraudulent clicks. Users with low trust scores have their clicks filtered out before affecting item ratings, creating a protective cushion against fraud without requiring real-time detection of each fraudulent click.
2Productivity
If the system updates internal ratings immediately upon user clicks, then rating updates are fast and responsive, but fraudulent clicks can contaminate content and recommendation systems
Solution Approach 1:
The trust score acts as an intermediary between user clicks and rating updates. Instead of directly updating ratings from all clicks, the system uses trust scores as a mediating filter that determines whether clicks should contribute to rating updates, thereby protecting rating accuracy while maintaining update efficiency.
Solution Approach 2:
The system implements self-service by automatically calculating trust scores and filtering fraudulent clicks without requiring manual review of each click. The automated trust score mechanism serves itself to maintain rating integrity, reducing the need for human intervention while preserving rating accuracy.
3Measurement precision
If the system collects extensive user data to build comprehensive user models for fraud detection, then user trustworthiness can be accurately assessed, but system complexity and data processing requirements increase
Solution Approach 1:
The patent segments user data collection into distinct categories including user profiles, behavior patterns, historical interactions, and demographic information. This segmentation allows the system to build comprehensive user models through modular data collection and processing, managing complexity while maintaining measurement precision.
Solution Approach 2:
The user model serves multiple functions: it detects fraudulent clicks, calculates trust scores, personalizes recommendations, and prevents fraud. This multi-functionality justifies the complexity of data collection, as the same comprehensive user model provides multiple benefits beyond just fraud detection.
4Reliability
If the system requires approval from multiple trustworthy users before updating ratings, then fraudulent rating updates are prevented, but the rating update process becomes slower and more complex
Solution Approach 1:
The system uses feedback from multiple trustworthy users to validate rating updates. Before a rating is updated, the system collects feedback signals from other users with high trust scores, ensuring that fraudulent updates are detected and rejected. This feedback mechanism enhances security while maintaining a relatively simple process through automated validation.
Data Source
AI summary
A method for detecting click fraud. In one embodiment, the method includes the steps of: directing a user to an item but delaying the internal ratings update of the item; collecting data to build a model of the user; building the model of the user in response to the collected data; determining the trustworthiness of the user in response to the model of the user; and updating internal ratings of the item in response to the trustworthiness of the user. In various embodiments, the click fraud can be the inflation of advertisement clicks, the inflation of popularity or the inflation of reputation.