Meta-Model Ensemble for Clinical Prediction with Missing Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Machine learning models in clinical settings face limitations due to inter-patient variability in data availability, where many patients lack recorded findings for various data categories, leading to data heterogeneity and the inability to perform predictions when missing data categories are present.
Innovation Solution
A meta-model comprising multiple machine learning models, each trained on different subsets of data categories, is used to perform clinical predictions. This approach selects models based on the data categories present in patient data, combining their prediction results weighted by their performance metrics to generate a comprehensive and accurate prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single machine learning model is trained to use all required data categories for prediction, then prediction accuracy can be maximized when complete data is available, but the model becomes unable to perform predictions when certain data categories are missing
Solution Approach 1:
The patent divides a single comprehensive prediction model into multiple specialized machine learning models, where each model is trained on a specific subset of data categories. This segmentation allows each model to be optimized for particular data types while the ensemble handles various combinations of available data, resolving the contradiction between accuracy and adaptability to missing data.
Solution Approach 2:
The patent creates a universal prediction system through an ensemble of multiple machine learning models that can collectively handle any combination of available data categories. Each model serves multiple functions by being applicable to different patient data scenarios, allowing the system to adapt to varying data availability while maintaining prediction capability.
2Adaptability or versatility
If multiple machine learning models are used to handle different data categories, then the ability to make predictions with available data improves, but the complexity of the prediction system increases
Solution Approach 1:
The patent introduces an intermediary component that automatically selects and weights appropriate machine learning models based on the available patient data categories. This intermediary layer manages the complexity of coordinating multiple models by providing a systematic approach to model selection and ensemble aggregation, reducing the operational burden despite having multiple specialized models.
3Productivity
If data imputation is performed to fill missing data categories, then single models can make predictions, but the reliability of predictions may be compromised due to imputation errors
Solution Approach 1:
The patent extracts and removes the need for data imputation by designing specialized machine learning models that are trained to make accurate predictions using only the specific data categories that are actually available for each patient. This extraction of the imputation step eliminates the source of imputation errors while maintaining prediction capability through the ensemble approach.
Data Source
AI summary
A method comprises: receiving data corresponding to a plurality of data categories of a patient; selecting, from a plurality of trained machine learning models and based on the plurality of data categories, a first machine learning model and a second machine learning model, the first machine learning model being trained using first data of a first subset of the plurality of data categories and having a first weight indicative of a first performance metric value, the second machine learning model being trained using second data of a second subset of the plurality of data categories and having a second weight indicative of a second performance metric value; generating a first prediction result and a second prediction result using, respectively, the first model and the second model; and generating a combined prediction result based on the first prediction result, the second prediction result, the first weight and the second weight.