Cursor Path Vector Analysis for Click Fraud Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting click fraud, such as automated scripts using zombie machines, are difficult to distinguish from legitimate human clicks, leading to unethical and illegal charges for advertisers, with estimates suggesting fraudulent clicks represent 2-20% of all clicks.
Innovation Solution
The method involves tracking the movement of a cursor on a web page before an advertisement is clicked and mapping this movement to vectors, which are analyzed to determine if the click is human-like or automated, using stochastic patterns in cursor movement to establish whether a human or automated tool is involved, thereby detecting click fraud.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated scripts use zombie machines to simulate clicks, then the ability to generate clicks is improved, but the ability to distinguish fraudulent clicks from legitimate clicks deteriorates
Solution Approach 1:
The system performs preliminary analysis of cursor movement paths before registering a click as valid. By mapping the cursor trajectory to vectors and analyzing stochastic patterns in advance, the system can identify automated click patterns before they result in fraudulent charges, thus improving fraud detection without affecting legitimate click processing
Solution Approach 2:
The patent replaces traditional mechanical click detection (simply detecting mouse button presses) with a more sophisticated analysis system that examines cursor movement trajectories. By substituting simple position detection with vector-based path analysis, the system can distinguish between human-controlled cursors and automated script-generated clicks
2Reliability
If cursor movement tracking and vector mapping is implemented, then click fraud detection capability is improved, but system complexity increases
Solution Approach 1:
The cursor movement analysis is segmented into discrete vector segments between successive cursor positions. By breaking down the continuous cursor path into manageable vector components, the system can analyze stochastic patterns without requiring complex continuous trajectory modeling, thus improving detection reliability while maintaining manageable system complexity
Solution Approach 2:
The patent introduces vector mapping as an intermediary representation between raw cursor positions and fraud detection logic. By converting cursor coordinates into vector segments with stochastic pattern characteristics, the system creates an intermediate data structure that simplifies the detection algorithm while improving its ability to distinguish human from automated behavior
Data Source
AI summary
A system and method for detecting click fraud where the movement of a cursor on a web page is tracked prior to an advertisement on the web page being clicked by a controller of the cursor. The tracking is performed by client code on a client device displaying the web page. A path taken by the movement of the cursor is mapped to at least one vector associated with the path by the client code. The at least one vector is useable to detect a click fraud.