Outlier Detection Method Using Combined Difference Distribution

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing outlier detection methods for measured values in various applications are inefficient in real-time detection and require expert analysis and prior knowledge of data properties, making them time and cost-intensive.

Innovation Solution

A computer-implemented outlier detection method that continuously records data, filters measured values, and determines a combined distribution of differences and noise, allowing for autonomous, real-time outlier identification without requiring expert analysis or prior knowledge.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional outlier detection methods are used, then outlier detection accuracy can be improved, but expert analysis and prior knowledge are required, making the process time and cost intensive

Engineering Contradiction:

Improveoutlier detection accuracyVSAvoidtime and cost for parameter determination

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs self-adjustment by automatically determining optimal parameters through iterative optimization based on synthetic training data, eliminating the need for expert analysis and manual parameter tuning while maintaining high detection accuracy

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system performs preliminary training with synthetic data before actual outlier detection, pre-determining optimal parameters and creating a ready-to-use detection model that can operate autonomously without requiring expert intervention during real-time detection

Inventive Principle:

Principle #10Preliminary action

2Productivity

If real-time outlier detection is implemented, then productivity is improved, but accurate parameter determination becomes more difficult without prior knowledge of data properties

Engineering Contradiction:

Improvereal-time detection capabilityVSAvoidparameter determination difficulty

Core Design Contradiction:

ProductivityVSDifficulty of detecting and measuring

Solution Approach 1:

The system replaces manual expert analysis with an automated computational approach that uses iterative optimization algorithms to determine parameters, enabling real-time detection while eliminating the difficulty of manual parameter determination

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system dynamically adjusts detection parameters based on the specific properties of the input data through automated optimization, allowing real-time detection across different applications without requiring pre-knowledge of data characteristics

Inventive Principle:

Principle #35Parameter changes

3Measurement precision

If manual parameter adjustment is performed, then detection accuracy can be improved, but device complexity and expert involvement increase

Engineering Contradiction:

Improveoutlier detection accuracyVSAvoidcomplexity of parameter determination process

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system automatically determines and adjusts its own parameters through iterative optimization algorithms, eliminating the need for manual expert adjustment while maintaining high detection accuracy and reducing operational complexity

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS20240019268A1Outlier detection method of detecting outliers in measured values of a measurand

Publication Date: 2024.01.18 ENDRESSHAUSER GRP SERVICES AG

AI summary

A method of detecting outliers in measured values of a measurand is disclosed, comprising the steps of: based on training data determining a combined distribution of differences between individual measured values and the filtered value of the measured value preceding the respective individual measured value to be expected in the application where the method is applied based on difference distribution of first differences of the filtered values of the measured values and a noise distribution of noise included in the measured values. Next, new measured values are identified as outliers when a probability of occurrence of a difference between the respective new measured value and the filtered value of the preceding measured value according to the combined distribution is lower than a predetermined level of confidence.