Hybrid Prediction Scores for Temporal Data Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing predictive data analysis systems face challenges in accurately processing temporal relationships across a large number of prediction input codes, leading to inefficiencies and reliability issues due to limited labeled data and missing medical health records, which affect the performance of disease progression models.
Innovation Solution
The system employs hybrid prediction scores combining co-occurrence-based and temporal prediction scores, using co-occurrence-based historical and temporal representations to improve predictive data analysis accuracy, by integrating co-occurrence-based and temporal machine learning models to generate hybrid scores for predictive entities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing machine learning systems are used for predictive data analysis with temporal relationships across large numbers of prediction input codes, then the system can process data, but the accuracy is insufficient due to limited labeled data and missing medical health records
Solution Approach 1:
The patent combines co-occurrence-based machine learning models with temporal machine learning models to create a hybrid system. The co-occurrence model processes co-occurrence-based historical representations of prediction input codes, while the temporal model processes temporal historical representations with timestamp information. Their prediction scores are integrated to produce hybrid prediction scores, thereby improving accuracy without requiring additional labeled data or complete medical records.
2Productivity
If traditional predictive analysis methods are used, then the system can operate with available data, but computational efficiency is low and throughput is reduced
Solution Approach 1:
The patent segments the predictive analysis task into two independent parallel processes: one processing co-occurrence-based historical representations through a co-occurrence model, and another processing temporal historical representations through a temporal model. This segmentation allows both models to operate simultaneously and independently, reducing overall computational time while maintaining high throughput.
3Measurement precision
If complete medical health records are required for accurate prediction, then prediction accuracy would improve, but the system becomes inoperable with incomplete or missing records
Solution Approach 1:
The patent introduces co-occurrence-based historical representations as an intermediary that captures relationships between prediction input codes without requiring complete medical health records. This intermediary representation, combined with temporal representations, allows the system to achieve accurate predictions even when medical records are incomplete or missing, thereby increasing adaptability to real-world data conditions.
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 with respect to input data entities that describe temporal relationships across a large number of prediction input codes. Certain embodiments of the present invention utilize systems, methods, and computer program products that perform predictive data analysis by using hybrid prediction scores that are determined based at least in part on co-occurrence-based prediction scores and temporal prediction scores, where the co-occurrence-based prediction scores are determined based at least in part on co-occurrence-based historical representation of a sequence of prediction input codes and temporal historical representation of the sequence of prediction input codes.