Blockchain Transaction Decomposition for Fraud Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Analyzing and visualizing fraudulent transactions in blockchain-based cryptocurrencies is challenging due to their complex and voluminous nature, with each transaction potentially involving thousands of inputs and outputs, making real-time identification and labeling cumbersome.
Innovation Solution
A computer-implemented method that decomposes blockchain transactions into a table format, aggregates entity and transaction information, and performs outer joins to create a block analysis table, enabling efficient analysis and visualization of transactions, including labeling fraudulent activities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional heuristic methods are used to analyze and label transactions as fraudulent, then transaction analysis can be performed, but the process becomes computationally expensive and difficult to achieve in real time due to the voluminous and complex nature of blockchain data with up to 2000 inputs and outputs per transaction
Solution Approach 1:
The patent segments the complex blockchain analysis task into distinct components: data extraction from blocks, entity information aggregation, transaction information aggregation, and fraudulent activity detection. By dividing the voluminous blockchain data into manageable segments and processing them through separate aggregation functions, the system reduces computational complexity while maintaining fraud detection accuracy
Solution Approach 2:
The patent introduces an intermediary aggregation layer that processes raw blockchain data before fraud detection analysis. The entity information aggregation and transaction information aggregation functions act as intermediaries, transforming the complex 2000+ input/output data into condensed representations that can be analyzed in real-time without losing critical fraud detection capabilities
2Reliability
If detailed analysis of all inputs and outputs (up to 2000 per transaction) is performed to identify fraudulent activity, then comprehensive fraud detection is achieved, but the computational cost becomes prohibitively expensive
Solution Approach 1:
The patent extracts only the essential information needed for fraud detection from the voluminous blockchain data. The aggregation functions selectively extract and consolidate critical entity and transaction attributes, discarding redundant data while preserving the reliability of fraud detection by focusing on the most indicative features
Solution Approach 2:
The patent applies partial action by performing aggregation on a subset of the most relevant transaction features rather than analyzing all 2000+ inputs and outputs in detail. This selective aggregation approach maintains sufficient fraud detection reliability while significantly reducing computational cost and energy consumption
3Difficulty of detecting and measuring
If comprehensive transaction data with 100+ features per input and output is analyzed, then thorough fraud investigation is possible, but the analysis pathway becomes cumbersome and difficult to visualize
Solution Approach 1:
The patent merges the 100+ features per input and output into consolidated aggregation results through entity information aggregation and transaction information aggregation. This combining process transforms the cumbersome detailed data into a unified, simplified view that maintains fraud detection capability while greatly improving ease of operation and visualization
Data Source
Figure 1~2
Figure 3
Figure 4A~4B
AI summary
A computer implemented method of decomposing a blockchain comprising transactions in digital currency for analysis and display is described. The method comprises first determining a range of blocks in the blockchain. Each block in the range of blocks in the blockchain is then unpacked into a table comprising one or more rows of input and output data for each transaction stored in the block. For the range of blocks in the blockchain, entity information and transaction information are then aggregated into a block analysis table. A node of a computing network and a computer program product adapted for implementation of such a method are also described.