Parameter Selection for Patient Prognosis Prediction Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing patient prognosis prediction models using machine learning face challenges due to varying parameter frequencies and combinations in clinical data, leading to reduced prediction accuracy and insufficient training data quality when parameters are missing.
Innovation Solution
An information processing device and method that calculates acquisition rate and frequency of parameters in patient time-series data, selecting high-quality parameters for training data based on these metrics to enhance data quality and quantity for machine learning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If only data from patients with high parameter satisfaction is used for training, then the quality of training data is improved, but the number of available training data sets is reduced
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the parameter selection criteria based on data characteristics. It calculates acquisition rates and frequencies for multiple parameters and selectively includes parameters that meet predetermined thresholds, thereby maintaining high data quality while expanding the pool of usable training data through flexible parameter inclusion strategies.
2Manufacturing precision
If parameters are selected based on high acquisition rate and frequency, then the quality of training data is improved, but the number of parameters that can be acquired is reduced
Solution Approach 1:
The patent implements partial action by selectively including only those parameters that meet predetermined acquisition rate and frequency thresholds. Rather than requiring all parameters to satisfy strict criteria, the system partially includes parameters based on their individual performance metrics, thereby maintaining data quality while preserving parameter variety through selective inclusion.
3Productivity
If machine learning is performed with insufficient or low-quality training data, then the prediction accuracy of the prognosis model is reduced
Solution Approach 1:
The patent applies feedback by calculating acquisition rates and frequencies for each parameter and using these metrics to inform the parameter selection process. The system establishes predetermined thresholds based on data characteristics and uses the calculated metrics to determine which parameters meet the quality standards, creating a feedback loop that ensures only high-quality parameters are included in the training data, thereby improving both prediction accuracy and model reliability.
Data Source
AI summary
An information processing device used in a system that predicts a prognosis of a patient by machine learning, the information processing device including: an input unit that receives an input of a plurality of sets of time-series data corresponding to a plurality of patients, the time-series data including a plurality of first parameters related to at least one of a condition and a treatment of each of the patients; and a processing unit that calculates an acquisition rate and an acquisition frequency of each of the first parameters included in the plurality of sets of time-series data, and selects a second parameter to be used for training data from the plurality of first parameters by using at least one of the calculated acquisition rate and the calculated acquisition frequency.