Column Classification Models for PHI Anomaly Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing classification-based predictive data analysis solutions face efficiency and reliability challenges when processing named data collections, particularly in identifying Protected Health Information (PHI) and performing column classification and anomaly detection, due to limitations in handling varied data formats and missing metadata.

Innovation Solution

The implementation of machine-learning-based techniques that generate varied training data through column augmentation and sub-column sampling, allowing column classification machine learning models to perform both classification and anomaly detection efficiently, and a deep learning framework for automatic PHI identification and data anonymization.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If existing classification-based predictive data analysis solutions are used to process named data collections, then basic classification functionality is provided, but efficiency and reliability are insufficient when handling varied data formats and missing metadata

Engineering Contradiction:

ImprovereliabilityVSAvoidhandling varied data formats

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent changes the parameters of the machine learning model by generating multiple sample sequences with different sample sizes from the same training column. This allows the model to learn robust classification patterns that are invariant to variations in data format and length, thereby improving reliability while maintaining adaptability to varied data formats.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent performs preliminary data preparation by generating multiple sample sequences and permuting them before training the machine learning model. This preliminary action ensures that the model is exposed to diverse data configurations in advance, enabling it to handle varied data formats and missing metadata reliably during inference.

Inventive Principle:

Principle #10Preliminary action

2Reliability

If machine learning models are trained with multiple sample sequences and permutations to improve robustness, then reliability and anomaly detection capability are enhanced, but training time and computational resources increase

Engineering Contradiction:

ImproverobustnessVSAvoidtraining time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent applies partial action by generating multiple sample sequences with different sample sizes rather than using all possible permutations. This selective approach provides sufficient diversity for robust training while avoiding the excessive computational burden of exhaustive permutation, thereby reducing training time while maintaining reliability.

Inventive Principle:

Principle #16Partial or excessive action

3Reliability

If column augmentation and sub-column sampling are used to generate varied training data, then model robustness and anomaly detection performance improve, but data processing complexity increases

Engineering Contradiction:

Improvemodel robustnessVSAvoiddata processing complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the training column into multiple sample sequences with different sample sizes. This segmentation creates diverse training inputs that improve model robustness while maintaining manageable processing complexity through systematic subsampling rather than complex data transformation.

Inventive Principle:

Principle #1Segmentation

4Productivity

If manual PHI identification and data anonymization processes are used, then data privacy protection is achieved, but processing efficiency and accuracy are reduced

Engineering Contradiction:

Improveprocessing efficiencyVSAvoidaccuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent replaces manual PHI identification and anonymization processes with a machine learning-based automated system. The trained column classification model automatically identifies PHI columns and triggers anonymization, substituting mechanical manual processes with intelligent automated processing that improves both efficiency and accuracy.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables self-service by allowing the machine learning model to automatically identify PHI columns and initiate anonymization without human intervention. The model serves itself by detecting sensitive data patterns and triggering the appropriate privacy protection actions autonomously.

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS20220222570A1Column classification machine learning models

Publication Date: 2022.07.14 OPTUM TECH INC

AI summary

There is a need for more effective and efficient performing classification-based predictive data analysis on named data collections. This need can be addressed by, for example, solutions for performing classification-based predictive data analysis on named data collections that utilize at least one of techniques for generating column classification machine learning models to perform column classification, techniques for generating column classification machine learning models to perform anomaly detection, techniques for utilizing trained column classification machine learning models to perform column classification, and techniques for utilizing trained column classification machine learning models to perform anomaly detection.