Anomaly Scoring Model for Subtle Physical Condition Changes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies fail to detect small changes in a monitored person's physical condition that may lead to an anomaly, increasing the risk of severe illness due to overlooked signs.
Innovation Solution
A trained model generation method using natural language processing and supervised learning to analyze care records and activity data, generating a model that outputs an anomaly score indicating the degree of physical condition anomalies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a notification system is implemented only when vital information indicates an anomalous value, then false alarms are reduced, but small changes in physical condition that may lead to anomalies are missed
Solution Approach 1:
The system performs preliminary analysis by extracting features from activity data and using a natural language processing model to determine anomaly presence before generating final notifications. This preliminary action enables early detection of small physical changes that precede actual anomalies, allowing the system to maintain high reliability while improving measurement precision of subtle condition changes.
Solution Approach 2:
The patent introduces an intermediary natural language processing model that analyzes care records and activity data to bridge the gap between raw sensor data and anomaly detection. This intermediary layer processes text data and activity features to generate deterministic anomaly presence information, enabling the system to detect small physical changes while maintaining reliable anomaly notification.
2Measurement precision
If manual monitoring of physical condition changes is performed, then detection accuracy is improved, but caregiver workload increases
Solution Approach 1:
The system implements self-service by automatically extracting features from activity data and using the natural language processing model to determine anomaly presence without requiring manual caregiver analysis. The system serves itself by processing care records and generating anomaly determinations autonomously, thereby improving measurement precision while reducing caregiver workload.
Solution Approach 2:
The patent replaces the mechanical system of manual monitoring with an automated information processing system. The natural language processing model and feature extraction algorithms substitute human caregivers' manual analysis, maintaining high measurement precision for physical condition monitoring while significantly easing operational burden on caregivers.
3Measurement precision
If comprehensive activity data analysis is performed for every subject, then detection accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The system extracts only the necessary features from comprehensive activity data using the natural language processing model, rather than analyzing all raw data. This extraction approach maintains high measurement precision for anomaly detection by focusing on relevant features while significantly reducing processing time and computational resource requirements.
Solution Approach 2:
The patent applies partial action by performing deterministic anomaly presence determination only for specific features extracted from activity data, rather than comprehensively analyzing all possible data points. This selective approach maintains sufficient detection precision while minimizing processing time and resource consumption.
Data Source
AI summary
A trained model generation method includes: determining, per first period, whether a subject has an anomaly in a physical condition, based on a care record including text data; extracting, based on activity data on the subject in a second period that includes a plurality of first periods each being the first period, one or more features calculated for one or more first periods in which the subject is determined to have no anomaly in the physical condition, the one or more features being extracted from among one or more features per first period; and generating, through learning, a trained model that uses a feature of the subject as an input and outputs an anomaly score indicating a degree of an anomaly in the physical condition of the subject, the learning using the one or more features as input data, and using the anomaly score as training data.