Feature Descriptor Generation for Interpretable Predictive Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In predictive analysis, as the number of generated features increases, it becomes difficult for data analysts to understand the meaning of these features, necessitating a system that can provide easily understandable descriptions of features while maintaining efficient generation.
Innovation Solution
An information processing system that includes a descriptor generation unit, a feature generation unit, and a feature explanation generation unit, which use predefined templates to generate feature descriptors and explanations from input tables, allowing for the creation of feature explanations that can be easily understood by humans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the number of generated features is increased, then the possibility of finding useful knowledge for prediction is improved, but the difficulty of understanding the meanings of features increases
Solution Approach 1:
The patent introduces feature descriptors as intermediary elements that bridge the gap between generated features and human understanding. These descriptors provide natural language explanations that make complex features interpretable, allowing data analysts to understand feature meanings without reducing the number of generated features.
Solution Approach 2:
The patent segments the feature generation process into distinct components: feature generation, descriptor generation, and explanation generation. This segmentation allows each component to be optimized independently, enabling the system to generate many features while maintaining interpretability through dedicated explanation mechanisms.
2Adaptability or versatility
If manual trial-and-error operations are used for pre-processing, then flexibility in designing procedures is improved, but the efficiency and productivity decrease
Solution Approach 1:
The patent implements automated feature generation and description systems that perform pre-processing tasks without requiring manual trial-and-error operations. The system autonomously generates features, creates descriptors, and provides explanations, thereby improving productivity while maintaining flexibility through programmable parameters and templates.
Solution Approach 2:
The patent performs feature generation and descriptor creation as preliminary actions before the actual predictive analysis. By preparing features and their descriptions in advance through automated processes, the system eliminates the need for manual pre-processing while ensuring that all necessary features are ready for analysis.
3Ease of operation
If templates are used to generate feature descriptors and explanations, then the ease of understanding features is improved, but the device complexity increases
Solution Approach 1:
The patent uses templates with configurable parameters to generate feature descriptors and explanations. By changing the parameters of existing templates rather than creating new complex systems, the patent achieves flexible and understandable feature descriptions while keeping the underlying system structure relatively simple and maintainable.
Data Source
AI summary
A descriptor generation unit 81 uses a first template prepared in advance to generate a feature descriptor, which generates a feature that may affect a prediction target from a first table including a variable of the prediction target and a second table. A feature generation unit 82 generates the feature by applying the feature descriptor to the first and second tables. A feature explanation generation unit 83 generates a feature explanation about the feature descriptor or the feature on the basis of a second template. An accepting unit 84 accepts values to be assigned to the first and second templates. The descriptor generation unit 81 generates the feature descriptor by assigning the accepted values to the first template, and the feature explanation generation unit 83 generates the feature explanation by assigning the values assigned to the first template to the second template.