Graph-Based NCD Risk Prediction Using Person-to-Person Distance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods fail to effectively predict non-communicable diseases (NCDs) by considering person-to-person spread of risk factors, which are not traditionally viewed as infectious but can influence disease development through social ties.
Innovation Solution
A computer-implemented method using graph-based machine learning with metric learning and regularization to detect and compute person-to-person distances based on risk factors associated with NCDs, represented by an adjacency matrix, to predict disease risk.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional disease prediction methods are used, then prediction simplicity is maintained, but prediction accuracy deteriorates due to ignoring person-to-person spread of risk factors
Solution Approach 1:
The patent segments the prediction problem by separating individual risk factors from social network structures. It uses graph representation to divide the complex interaction between individuals and their social environments into manageable components (nodes for individuals, edges for relationships), enabling accurate prediction while maintaining computational tractability.
Solution Approach 2:
The patent introduces a new dimension by transforming traditional scalar risk factor data into graph-structured data. By representing risk factors as edges in a graph connecting individuals, it adds a social network dimension to traditional prediction models, improving accuracy without overwhelming complexity through structured data representation.
2Measurement precision
If graph-based machine learning with metric learning is applied, then prediction accuracy improves, but computational complexity increases
Solution Approach 1:
The patent performs preliminary actions by pre-computing graph representations and metric learning parameters before actual prediction. It pre-processes social network data into structured graph formats and pre-calculates distance metrics between individuals, reducing computational burden during prediction while maintaining high accuracy.
Solution Approach 2:
The patent changes parameters by transforming raw social network data into standardized graph representations with specific metric parameters. It standardizes relationship types, individual attributes, and distance calculations into consistent parameter formats, enabling accurate predictions while simplifying computational processing through parameter standardization.
3Measurement precision
If risk factors are modeled as infectious spread through social networks, then prediction accuracy improves, but data requirements and processing complexity increase
Solution Approach 1:
The patent extracts only the essential information needed for prediction from vast social network data. It extracts key risk factor transmissions and critical social connections, filtering out redundant data while maintaining the core predictive signal. This selective extraction reduces data quantity requirements while preserving prediction accuracy.
Solution Approach 2:
The patent applies partial action by focusing graph-based analysis only on specific subgraphs relevant to NCD prediction, rather than processing entire social networks. It identifies and analyzes only the portions of the network where risk factor transmission is most likely, reducing data processing requirements while maintaining predictive accuracy for at-risk individuals.
Data Source
AI summary
A computer-implemented method for predicting non-communicable diseases with infectious risk factors using artificial intelligence includes detecting one or more risk factors associated with a non-communicable disease based on a graph associated with person-to-person links, generating a data structure for compactly representing the graph to compute at least one person-to-person distance, and performing a machine learning technique with regularization of the at least one person-to-person distance.