Hierarchical Classification via Online and Co-occurrence Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing classification systems are inefficient and unreliable in performing classification tasks in hierarchical prediction domains due to their inability to effectively handle hierarchical predictive relationships and the integration of structured and unstructured data.
Innovation Solution
The use of online machine learning, co-occurrence analysis, and fusion of structurally hierarchical and non-hierarchical predictions, along with Human Phenotype Ontology (HPO) predictions, to generate accurate and reliable predictive outputs in hierarchical prediction domains.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing classification systems are used for hierarchical prediction domains, then the system structure is simple, but the classification efficiency and reliability are poor
Solution Approach 1:
The patent segments the classification system into multiple specialized components: hierarchical prediction module, co-occurrence analysis module, structured fusion machine learning model, and non-structured fusion machine learning model. Each module handles specific aspects of the classification task, improving reliability through specialized processing while managing complexity through modular architecture.
Solution Approach 2:
The patent merges multiple prediction approaches (hierarchical predictions, co-occurrence analysis predictions, structured data predictions, and non-structured data predictions) into a unified classification system. The structured fusion model and non-structured fusion model combine these different prediction types to generate final classifications, improving reliability through ensemble methods.
2Measurement precision
If existing classification systems are used for hierarchical prediction domains, then the implementation is straightforward, but the predictive accuracy is poor
Solution Approach 1:
The patent introduces intermediary processing layers including the structured fusion machine learning model and non-structured fusion machine learning model. These intermediaries integrate predictions from multiple sources (hierarchical predictions, co-occurrence analysis, structured data, non-structured data) before generating final classifications, thereby improving predictive accuracy through multi-stage processing.
Solution Approach 2:
The patent employs composite prediction approaches by combining multiple types of predictions (hierarchical, co-occurrence, structured data-based, non-structured data-based) within fusion models. This composite strategy leverages the strengths of different prediction methods to achieve higher overall accuracy than any single method alone.
3Reliability
If hierarchical relationships are properly handled, then the classification reliability improves, but the computational complexity increases
Solution Approach 1:
The patent segments the handling of hierarchical relationships into a dedicated hierarchical prediction module that processes hierarchical data structures separately. This module generates hierarchical predictions that are then integrated by the structured fusion model, allowing complex hierarchical processing to be isolated and managed independently, improving reliability while controlling overall computational complexity.
Data Source
AI summary
There is a need for solutions that classification solutions in hierarchical prediction domains. This need can be addressed by, for example, performing one or more online machine learning, co-occurrence analysis machine learning, structured fusion machine learning, and unstructured fusion machine learning. In one example, structured predictions inputs are processed in accordance with an online machine learning analysis to generate structurally hierarchical predictions and in accordance with a co-occurrence analysis machine learning analysis to generate structurally non-hierarchical predictions. Then, the structurally hierarchical predictions and the structurally non-hierarchical predictions in accordance with processed by a structured fusion model to generate structure-based predictions. Afterward, the structure-based predictions and non-structure-based predictions are processed in accordance with an unstructured fusion model to generate one or more unstructured-fused predictions.