ML Model Generation Using Entropy-Based Attribute Correlation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing AI/ML models for AIOps tasks face challenges in identifying the most relevant subset of attributes for accurate model training, as humans struggle to determine which attributes are most useful without significant trial and error, and existing systems require manual input from subject matter experts.

Innovation Solution

An automated system that uses entropy computation and correlation analysis to identify the most relevant attributes for AIOps tasks, determining a subset of attributes with entropy values within a predefined range and computing correlations to generate ML models without human intervention.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual selection of attributes by subject matter experts is used, then domain knowledge can be leveraged, but the process requires significant time and human intervention

Engineering Contradiction:

Improveattribute selection accuracyVSAvoidmodel development time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs automated attribute selection using entropy computation and correlation analysis, eliminating the need for manual expert intervention. The algorithm independently identifies relevant attributes by computing entropy values and correlation coefficients, thereby reducing time loss while maintaining selection quality through objective mathematical criteria.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The manual mechanical process of expert attribute selection is replaced with an automated computational system. The system uses entropy functions and correlation calculations to objectively determine attribute relevance, substituting human expert judgment with a deterministic mathematical approach that operates automatically without human intervention.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If all candidate attributes are used for model training, then comprehensive data is available, but the model training complexity and computational cost increase

Engineering Contradiction:

Improvemodel accuracyVSAvoidmodel training complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system extracts only the most relevant attributes from the complete candidate set by computing entropy values and correlation coefficients. Attributes with entropy outside the predefined range or with low correlation to the target are extracted and excluded from training, leaving only the essential subset that maintains model accuracy while reducing complexity.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

Different attributes are evaluated individually using entropy computation to determine their information content. Attributes are selectively retained or excluded based on their local quality metrics (entropy values and correlation coefficients), rather than treating all attributes uniformly. This localized evaluation enables precise attribute selection that optimizes model training efficiency.

Inventive Principle:

Principle #3Local quality

3Extent of automation

If automated attribute selection using entropy and correlation is implemented, then manual intervention is reduced, but the computational overhead for entropy computation and correlation analysis increases

Engineering Contradiction:

Improveattribute selection automationVSAvoidcomputational resource consumption

Core Design Contradiction:

Extent of automationVSUse of energy by moving object

Solution Approach 1:

The system computes entropy and correlation for all candidate attributes initially (excessive action), but then uses the predefined entropy range filter to eliminate unnecessary computations for attributes outside the optimal range. This partial computation approach reduces overall computational overhead by avoiding correlation analysis for attributes that fail the entropy criterion, thereby balancing automation with resource efficiency.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20260065142A1Automated machine learning model generation using correlation between attributes

Publication Date: 2026.03.05 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

An example operation may include one or more of accessing table data including columns corresponding to candidate attributes, respectively, and a target column corresponding to a target attribute, determining, by executing an entropy function, entropy values of the candidate attributes based on randomness of values in the columns, respectively, identifying a subset of candidate attributes that have entropy values between a predefined range of entropy values, determining a correlation between the subset of attributes by executing a correlation function on values in columns, determining at least one machine learning (ML) model to be generated based on the correlation between the subset of candidate attributes, and training the at least one ML model to determine the target attribute based on execution of the at least one ML model on training data to generate a trained at least one ML model.