Probabilistic Clustering Rules Engine for Interpretable Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing predictive data analysis systems face challenges in computational efficiency and interpretability when generating rules engines, making them costly and difficult to implement.
Innovation Solution
The use of probabilistic clustering techniques to generate augmented clusters, including general, high-confidence, and low-confidence clusters, which improve computational efficiency and interpretability by capturing variations across prediction input data objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional predictive data analysis systems are used to generate rules engines, then comprehensive analysis can be performed, but computational cost and complexity increase significantly
Solution Approach 1:
The patent segments the data analysis process into multiple independent clustering operations. Instead of performing one comprehensive complex analysis, the system divides the data into different clusters and performs separate clustering operations on each segment. This segmentation reduces the computational complexity of each individual operation while maintaining overall analysis accuracy through the combination of results from multiple clusters.
2Measurement precision
If more rules are added to improve predictive accuracy, then analysis quality improves, but computational efficiency decreases
Solution Approach 1:
The patent applies partial action by performing clustering operations selectively on different data segments rather than applying the same comprehensive analysis to all data uniformly. Each cluster receives tailored clustering operations appropriate to its characteristics, reducing unnecessary computational effort on segments where simpler analysis suffices, while still achieving high predictive accuracy where needed.
3Reliability
If comprehensive clustering is performed on all data objects, then complete coverage is achieved, but processing time increases
Solution Approach 1:
The patent segments the comprehensive data set into multiple smaller clusters that can be processed independently and in parallel. This segmentation maintains complete coverage of all data objects while reducing the processing time for each individual cluster. The independent nature of cluster processing allows for parallel execution, further reducing total processing time while ensuring all data objects are analyzed.
Data Source
AI summary
Various embodiments of the present invention provide methods, apparatus, systems, computing devices, computing entities, and/or the like for performing predictive data analysis. Certain embodiments of the present invention utilize systems, methods, and computer program products that perform predictive data analysis by generating input processing rules using at least one of general clusters generated using all of a set of prediction input data objects, high-confidence clusters generated using prediction input data objects having threshold-satisfying clustering confidence scores, and low-confidence clusters generated using prediction input data objects having non-threshold-satisfying clustering confidence scores.