Fraud Detection System Using Blacklist and ML Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online ad-exchanges face significant challenges in detecting fraudulent activities, particularly botnets that simulate user traffic to inflate ad revenue, which leads to financial losses for advertisers and publishers.
Innovation Solution
A fraud detection system that extracts identifiers from advertisement requests, checks them against a blacklist, and uses machine learning models to determine whether to include sources in the blacklist based on features extracted from requests, thereby denying fraudulent requests and preventing botnet activity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a fraud detection system uses machine learning models to analyze advertisement requests, then the accuracy of detecting fraudulent activities improves, but the system complexity and processing time increase
Solution Approach 1:
The fraud detection system is divided into multiple independent components: a request handling unit that extracts identifiers and checks blacklists, and a model generation unit that trains machine learning models. This segmentation allows each component to specialize in specific tasks, improving detection accuracy while managing system complexity through modular design.
Solution Approach 2:
The system performs preliminary actions by maintaining a blacklist of known fraudulent identifiers and checking requests against it before applying complex machine learning analysis. This preliminary filtering reduces the number of requests requiring full model analysis, thereby improving detection accuracy for known threats while reducing overall system complexity and processing load.
2Reliability
If the system checks every identifier against a blacklist, then fraudulent requests are blocked more effectively, but the processing speed decreases
Solution Approach 1:
The system applies different levels of inspection to different requests based on their characteristics. Identifiers matching the blacklist receive immediate blocking (high scrutiny), while other requests undergo feature extraction and model analysis (moderate scrutiny). This local differentiation of quality control ensures reliable fraud prevention for known threats while maintaining processing speed for legitimate requests.
3Adaptability or versatility
If the system extracts and analyzes features from all requests, then the ability to detect new fraudulent patterns improves, but the computational resources required increase
Solution Approach 1:
The system performs partial feature extraction and analysis only for requests that pass the initial blacklist check or exhibit suspicious characteristics. Not all requests undergo full feature extraction and model analysis, which reduces computational resource consumption while maintaining the ability to detect new fraudulent patterns in the subset of requests that require detailed examination.
Data Source
AI summary
The present teaching relates to a fraud detecting system and method for providing protection against fraudulent advertisement requests. Upon receiving a request for an advertisement, the system extracts an identifier, associated with a source from which the request originates, included in the request. The system determines whether the extracted identifier is included in a list of designated identifiers, and when the identifier is included in the list, the system denies the request for the advertisement. When the identifier is not included in the list of designated identifiers, the system provides the advertisement in response to the request, and extracts a set of features from the request and other requests that originate from the source to determine whether the identifier associated with the source is to be included in the list of designated identifiers based on the set of features in accordance with one or more models.