Psychological Stress Prediction Using Similar Worker Group Models

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing systems fail to accurately predict the future level of psychological stress in individuals, which is crucial for effective depression prevention.

Innovation Solution

A prediction system that utilizes machine learning to generate models based on working style, attendance status, activity levels, and psychological stress data of similar worker groups to forecast future psychological stress levels.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If a prediction model is generated using general worker data without segmentation, then the system complexity is reduced, but the prediction accuracy of future psychological stress levels deteriorates

Engineering Contradiction:
Improvesystem complexityVSAvoidprediction accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent segments the worker population into similar worker groups based on working style similarity (e.g., teleworking rate, work hours) and attendance status similarity. This segmentation allows the prediction model to focus on homogeneous groups, improving prediction accuracy for future psychological stress levels while managing complexity through targeted data processing for each group.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If multiple types of worker information (working style, attendance, measurement, action records) are integrated into the prediction model, then the prediction accuracy improves, but the data processing complexity increases

Engineering Contradiction:
Improveprediction accuracyVSAvoiddata processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent processes different types of worker information separately by creating similar worker groups based on specific criteria (working style, attendance status), then integrates this segmented data into the prediction model. This approach manages data processing complexity by organizing information into structured groups before analysis.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transforms various worker information types into standardized parameters suitable for machine learning processing. Working style information (teleworking rate, work hours), attendance status, measurement information (activity levels, sleeping hours), and action record information are converted into quantifiable parameters that can be efficiently processed by the prediction model.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If similar worker groups are extracted based on detailed similarity criteria, then the prediction accuracy improves, but the time required for data processing increases

Engineering Contradiction:
Improveprediction accuracyVSAvoiddata processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent segments workers into similar worker groups using predefined similarity criteria (e.g., teleworking rate thresholds, work hour ranges, attendance patterns). This segmentation approach balances detail and efficiency by using practical grouping criteria that capture meaningful similarities without requiring excessive computational resources for continuous comparison.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4693127A1Prediction system, prediction method, and program
Publication Date: 2026.02.11 SHIONOGI & CO LTD
  • EP4693127A1 patent drawingFigure 1
  • EP4693127A1 patent drawingFigure 2
  • EP4693127A1 patent drawingFigure 3

AI summary

To accurately predict the level of future psychological stress of an object person. A prediction system (100) includes a first extraction unit (42) that extracts a similar worker group from a plurality of workers (W) on the basis of a degree of similarity of a working style to a working style of an object person (O) during a first period, a first model generation unit (43) that generates a first prediction model (57) by machine learning using first information, as learning data, the first information including at least working style information, measurement information, action record information, and psychological stress information of each worker (W) included in the similar worker group, and a first prediction unit (44) that predicts a level of psychological stress of the object person (O) by inputting input data including at least the working style information, the measurement information, and the action record information of the object person (O) to the first prediction model (57).