Lightning KNN Host System for Real-Time Transaction Data Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Processing and analyzing large volumes of transaction data is time-consuming and resource-intensive, hindering the identification and categorization of transaction data, which is crucial for transaction card issuers.
Innovation Solution
The Lightning KNN (known nearest neighbor) host system, which includes a historical data retrieval engine, distance evaluator, outer radius boundary determiner, and communication bus, enables efficient data processing by loading historical data, evaluating distances, and grouping records into sets, thereby improving processing efficiency and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional data processing methods are used to analyze large volumes of transaction data, then comprehensive data analysis can be performed, but processing time increases and resources are consumed excessively
Solution Approach 1:
The patent segments the large transaction data into smaller batches or chunks that can be processed in parallel. The data processing system divides the comprehensive data set into multiple segments, processes them simultaneously using distributed computing resources, and then aggregates the results. This segmentation enables faster processing while maintaining analysis accuracy by distributing the computational load across multiple processors or nodes.
Solution Approach 2:
The patent applies preliminary actions by pre-processing and preparing data structures before the main analysis operation. Historical transaction data is pre-aggregated, pre-filtered, and organized into optimized data structures that enable faster querying and analysis during the actual processing phase. This preliminary preparation reduces the computational burden during real-time processing, thereby reducing processing time while maintaining analytical completeness.
2Measurement precision
If traditional data processing methods are used to analyze large volumes of transaction data, then comprehensive data analysis can be performed, but resource consumption increases
Solution Approach 1:
By segmenting the data processing task into smaller units that can be executed in parallel across multiple processors or computing nodes, the system distributes resource consumption rather than concentrating it. This allows for more efficient utilization of available computational resources, reducing overall energy consumption while maintaining the ability to perform comprehensive data analysis through aggregated results from multiple segments.
Solution Approach 2:
The patent applies partial action by processing data in selective batches or focusing computation on the most relevant data subsets rather than uniformly processing entire data sets. The system identifies and prioritizes high-value transaction data for detailed analysis while using summarization or sampling for less critical data, thereby reducing overall resource consumption while maintaining adequate analysis accuracy for decision-making purposes.
3Productivity
If ceiling number radius is reduced to process fewer records, then processing speed increases, but identification accuracy of nearest neighbors decreases
Solution Approach 1:
The system performs preliminary actions by pre-computing and storing distance metrics, similarity scores, or hierarchical cluster structures during offline data preparation phases. These pre-computed structures enable the system to quickly retrieve nearest neighbors within the ceiling radius without performing exhaustive calculations during real-time processing. This preliminary preparation maintains identification accuracy while achieving high processing speeds during operational phases.
Solution Approach 2:
The patent uses copying by creating and maintaining simplified representations or proxies of the full data set, such as summary statistics, clustered groupings, or approximate nearest neighbor structures. These copied representations allow for rapid querying within the ceiling radius constraint while preserving the essential characteristics needed for accurate nearest neighbor identification. The system can work with these lighter-weight copies during high-speed processing and reference the full data when higher accuracy is needed.
Data Source
AI summary
Systems and methods of improving the operation of a transaction network and transaction network devices is disclosed. A lightning KNN host may comprise various modules and engines as discussed herein wherein lookalike records may be identified whereby the speed of the lightning KNN network may be enhanced and the accuracy and precision of results improved whereby the transaction network more properly functions according to approved parameters.