Survey Diagnosis Using Autoencoders for Missing-Data Precision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing oriental medicine diagnosis methods using surveys face challenges in accurately determining health conditions due to difficulties in designing survey questions, limiting their diagnostic utility.
Innovation Solution
A survey-based diagnosis method utilizing autoencoder-based deep learning models to analyze response information from survey questions, including those for energy and blood generation, circulation, and balance adjustment functions, to generate state vectors and diagnose health conditions based on similarity and reconstruction loss values, with missing value correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If survey technique is used for oriental medicine diagnosis, then ease of operation is improved, but measurement precision deteriorates
Solution Approach 1:
The patent transforms survey response data into numerical parameters and uses deep learning models to process these parameters, converting qualitative survey responses into quantitative diagnostic measurements that improve precision while maintaining ease of operation
Solution Approach 2:
The patent replaces traditional mechanical survey analysis methods with deep learning-based artificial intelligence systems, using neural networks to automatically analyze survey responses and generate diagnostic results, thereby improving measurement precision without increasing operational complexity
2Measurement precision
If deep learning model is used to analyze survey responses, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent develops a universal deep learning platform that can handle multiple survey types and diagnostic scenarios with a single system architecture, reducing the need for separate complex systems for different diagnostic purposes
Solution Approach 2:
The patent uses autoencoder-based deep learning models that learn to compress and reconstruct survey response patterns, creating simplified representations of complex diagnostic information that reduce system complexity while maintaining diagnostic accuracy
3Measurement precision
If autoencoder-based deep learning model is used, then measurement precision is improved, but loss of information increases due to missing values
Solution Approach 1:
The patent implements iterative imputation using autoencoder-based deep learning where the model generates predictions for missing values, evaluates reconstruction loss, and refines predictions through feedback loops, progressively improving accuracy while recovering information that would otherwise be lost
Solution Approach 2:
The patent performs preliminary data processing and imputation before main diagnostic analysis, using autoencoder models to pre-fill missing survey responses and prepare complete datasets, thereby preventing information loss in subsequent diagnostic steps
Data Source
AI summary
A survey-based diagnosis method and a system therefor are provided. The survey-based diagnosis method according to several embodiments of the present disclosure enables a diagnosee to be diagnosed on the basis of response information about the diagnosee with respect to a plurality of survey questions. The plurality of survey questions can include questions for diagnosing an energy and blood generation function, an energy and blood circulation function, and an energy and blood balance adjustment function, and the health condition of the diagnosee can be accurately diagnosed by using these questions. In addition, a quick and convenient oriental medicine diagnosis service can be provided using a questionnaire technique.