Network Address Attribute Association via Data Filtering and Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods face challenges in associating accurate and reliable attributes with network addresses due to poor data quality, unobservable information, privacy constraints, proxy addresses, and mobile network devices, which hinder the effectiveness of services relying on location, time, intent, and identity data.
Innovation Solution
A method involving filtering, mapping, and processing techniques to enhance data quality, translate unobserved attributes, cluster network address observations, and propagate attributes to similar addresses, ensuring accurate associations between network addresses and attributes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If data is collected from multiple sources to form associations between network addresses and attributes, then the quantity of available information increases, but the data quality deteriorates due to noisy and untrustworthy data
Solution Approach 1:
The patent applies preliminary action by filtering and validating data before it is used to form associations. The system pre-processes incoming observations to remove noisy and untrustworthy data, ensuring that only high-quality data contributes to the associations between network addresses and attributes.
Solution Approach 2:
The patent introduces an intermediary filtering mechanism that mediates between raw data collection and association formation. This intermediary layer validates and cleanses data from multiple sources before it reaches the association engine, preventing noise from degrading overall data quality.
2Measurement precision
If direct observation of network addresses is performed to determine attributes, then measurement precision is improved, but reliability deteriorates when information cannot be directly observed or is prohibited due to privacy constraints
Solution Approach 1:
The patent uses copying by creating derived attributes that replicate the information value of unobservable or restricted attributes without directly observing them. The system infers attributes like location and intent from observable network behavior patterns, creating accurate copies of the desired information through indirect measurement.
Solution Approach 2:
The patent replaces direct mechanical observation with indirect inference mechanisms. Instead of directly observing protected attributes, the system substitutes mathematical modeling and pattern recognition to derive the same information from observable network traffic characteristics.
3Productivity
If associations are formed based on single network address observations, then the processing speed is improved, but reliability deteriorates due to insufficient information
Solution Approach 1:
The patent merges multiple observations of the same network address to form composite association profiles. By combining information from multiple time points and observation sources, the system creates more reliable associations while maintaining efficient processing through aggregated data structures.
Solution Approach 2:
The patent applies preliminary action by pre-aggregating observations into clustered groups before forming associations. This pre-processing organizes multiple observations into meaningful patterns, enabling reliable association formation without requiring real-time processing of every individual observation.
4Ease of operation
If network addresses are used to identify end devices, then the ease of operation is improved, but reliability deteriorates when proxy addresses or mobile devices are involved
Solution Approach 1:
The patent applies dynamics by creating adaptive association profiles that evolve with network address behavior over time. The system dynamically adjusts association confidence levels based on observed patterns, allowing it to handle mobile devices and proxy addresses by recognizing behavioral consistency rather than relying on static address-device mappings.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In various embodiment, techniques are provided for associating attributes with network addresses based on network address observations. The techniques may involve filtering network address observations to ensure data quality, mapping/translating associations with observed attributes to one or more attributes that are not directly observed, processing network address observations and/or network address to attribute associations by clustering/grouping, processing network address observations and/or network address to attribute associations to determine refined attributes based on one or more other attributes, and propagating network address to attribute associations between network addresses.