Feature Screening for Physiological State Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for determining a test subject's physiological state, such as obesity, rely on body mass index (BMI) without considering metabolic status, limiting accuracy in monitoring physiological conditions.
Innovation Solution
An electronic device and method for screening features that are significantly related to a specific physiological state by using a processor, storage medium, and transceiver to collect and analyze physiological data, generate feature subsets based on models like random forest or support vector machine, and select key features and their accompanied features based on relation indices and performance indices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional methods using BMI are used to determine physiological state, then the measurement process is simple and quick, but the measurement precision is insufficient because it does not consider metabolic status
Solution Approach 1:
The patent segments the complex task of physiological state determination by separating it into two parts: (1) using machine learning models to identify and select relevant metabolite features from a large pool of possible metabolites, and (2) using only the selected features for final diagnosis. This segmentation allows the system to handle complexity through intelligent feature selection rather than analyzing all possible metabolites, thus improving accuracy while managing system complexity.
Solution Approach 2:
The patent extracts only the most relevant metabolite features from the complete set of physiological data using machine learning models (random forest, support vector machine, etc.). By taking out and focusing only on the selected features that have high predictive value, the system achieves better measurement precision without needing to process all available data, effectively resolving the contradiction between accuracy and complexity.
2Measurement precision
If all metabolites are analyzed to determine physiological state, then measurement precision improves, but the quantity of data to process increases significantly
Solution Approach 1:
The patent applies the extraction principle by using machine learning models to identify and select only the most relevant metabolite features from the complete metabolite set. The models evaluate multiple metabolites and extract only those that have high predictive value for the target physiological state, thereby reducing the quantity of data to process while maintaining or improving measurement precision.
Solution Approach 2:
The patent applies partial action by analyzing only the selected subset of metabolites rather than all possible metabolites. The machine learning models determine the optimal subset size and composition, allowing the system to achieve sufficient precision with a manageable number of metabolites, avoiding the excessive data processing that would result from analyzing every possible metabolite.
3Measurement precision
If multiple metabolites are selected for analysis, then prediction accuracy improves, but the computational resources and time required increase
Solution Approach 1:
The patent applies preliminary action by pre-training machine learning models on large datasets to identify and select relevant metabolite features before actual physiological state determination. This preliminary feature selection process creates a streamlined set of metabolites to analyze, significantly reducing the computational time required during actual diagnosis while maintaining high prediction accuracy through the pre-identified relevant features.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An electronic device (100) and a method for screening features for predicting a physiological state are provided. The method includes: obtaining multiple physiological data corresponding to multiple features; generating multiple first subsets of the multiple features according to the multiple physiological data based on a first model, wherein the multiple first subsets respectively correspond to the multiple physiological data; selecting a first feature from the multiple features according to the multiple first subsets, calculating a first relation index of the first feature and a second feature corresponding to the multiple features, and selecting the second feature as an accompanied feature of the first feature according to the first relation index; and outputting the first feature and the accompanied feature.