Feature Trait Detection in Imbalanced Datasets via Regression Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current classification algorithms have a low recall rate and fail to accurately detect feature traits of a group, such as customers with high net worth, as they do not effectively focus on the accuracy of feature contribution in predicting outcomes.
Innovation Solution
A computer-implemented method involving building regression models to compute feature contribution scores, selecting features based on variance between positive and negative cases, and generating prediction probability values to identify feature traits from an imbalanced dataset, with a focus on predictor importance values exceeding threshold values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current classification algorithms are used on imbalanced datasets, then the processing speed is maintained, but the recall rate is low and feature trait detection accuracy deteriorates
Solution Approach 1:
The patent segments the feature selection process into multiple stages: first selecting features with high contribution scores, then filtering by variance thresholds, and finally ranking by predictor importance. This multi-stage segmentation allows the system to systematically identify relevant features from imbalanced datasets, improving both recall rate and detection accuracy without compromising processing efficiency.
Solution Approach 2:
The patent performs preliminary actions by pre-computing contribution scores for all features using regression models before the actual classification process. This preliminary feature scoring and filtering prepares the data in advance, ensuring that when classification occurs, the most relevant features are already identified, thereby improving recall rate and feature trait detection accuracy.
2Measurement precision
If traditional classification algorithms are applied, then computational resources are conserved, but the ability to accurately detect feature traits of partitioned groups deteriorates
Solution Approach 1:
The patent extracts only the most relevant features from the complete feature set by applying contribution score thresholds and variance filters. This extraction process removes irrelevant features before feeding data to the classification model, improving feature trait detection accuracy while actually reducing the computational burden and model complexity by working with a smaller, more focused feature subset.
Solution Approach 2:
The patent applies different quality criteria to different stages of feature selection: contribution scores measure feature relevance to the target, variance thresholds ensure feature discrimination between positive and negative cases, and predictor importance rankings prioritize the most influential features. This multi-criteria local quality assessment improves detection accuracy without requiring uniformly complex processing across all features.
3Measurement precision
If feature selection based on contribution scores and variance is implemented, then feature trait detection accuracy is improved, but the processing time increases
Solution Approach 1:
The patent applies partial action by selecting only a subset of features that meet the contribution score and variance thresholds, rather than processing all features. This partial feature selection achieves sufficient detection accuracy for the application while significantly reducing processing time compared to analyzing the complete feature set.
Data Source
AI summary
A computer-implemented method, system and computer program product for detecting feature traits from an imbalanced dataset. A first regression model is built based on a simulated dataset to compute contribution scores for the features to make a target a positive case. A variance in the features' original values for each feature of the first set of features (those features with contribution scores for positive cases that exceed a threshold value) between the positive and negative cases is determined. A second regression model is built to calculate a predictor importance value for a second set of features (features from the first set of features with a variance in their original values for both positive and negative cases that exceeds a threshold value). Feature traits are then extracted from a group of clustered positive cases with features of the second set of features containing a predictor importance value exceeding a threshold value.