Stream Data Window Specification Generation via Similarity Metrics

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for defining window specifications in complex event processing are prone to errors and complexity, making it difficult to accurately divide data streams for detecting abnormal behavior in scenarios like facility surveillance, leading to potential security holes.

Innovation Solution

A method that generates proposed window specifications based on a similarity metric to identify similar data items in the stream, allowing for the automatic detection of patterns and improving the accuracy of processing specifications, with user interaction for fine-tuning.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If manual processing specifications are defined using complex syntactical notations, then the system can process streaming data, but the specifications become error-prone and difficult to define correctly

Engineering Contradiction:

Improveease of defining processing specificationsVSAvoidaccuracy of window specifications

Core Design Contradiction:

Ease of operationVSReliability

Solution Approach 1:

The system automatically generates processing specifications by analyzing the data stream itself, allowing the data to define its own windows through pattern recognition rather than requiring manual specification. The system serves itself by extracting window definitions from the data characteristics.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces manual mechanical definition of window specifications with an automated algorithmic system that uses similarity metrics and pattern recognition to generate specifications, substituting human cognitive effort with computational analysis.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Adaptability or versatility

If complex syntactical notations are used to define windows, then flexible window definitions are possible, but the complexity increases making correct formulation difficult

Engineering Contradiction:

Improveflexibility of window definitionsVSAvoidcomplexity of processing specifications

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent extracts the essential characteristics of window definitions directly from the data stream by identifying patterns and similarities, separating the critical features from the complex syntactical framework of traditional query languages.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The system changes the parameters used to define windows from complex syntactical rules to simpler similarity metrics and pattern-based characteristics, transforming the definition mechanism while maintaining flexibility.

Inventive Principle:

Principle #35Parameter changes

3Productivity

If traditional window division methods are used, then processing can be performed, but incorrect division may occur leading to missed abnormal conditions

Engineering Contradiction:

Improveprocessing speedVSAvoidaccuracy of event detection

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system uses feedback from the data stream itself to continuously refine window definitions, analyzing patterns in the data to determine appropriate window boundaries and adjusting specifications based on observed data characteristics rather than fixed pre-defined rules.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent performs preliminary analysis of the data stream to identify patterns and characteristics before finalizing window specifications, preparing the processing framework in advance based on data inspection to ensure accurate event detection.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS9286353B2Method for generating processing specifications for a stream of data items

Publication Date: 2016.03.15 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

At least one processing specification is generated for a stream of data items captured by a sensor. A plurality of proposed window specifications is generated. The at least one processing specification is generated based on at least one of the proposed window specifications. The plurality of proposed window specifications being generated based on a similarity metric is configured to identify similar pairs of data items in the stream of data items.