Neural Network-Based Study about Correlation Model between TCM Constitution and Physical Examination Indexes Based on...

Area of Science:

• Integrative Medicine
• Computational Biology
• Traditional Chinese Medicine

Background:

• Current methods for identifying Traditional Chinese Medicine (TCM) constitution face challenges including a shortage of TCM doctors, complex processes, low efficiency, and limited clinical application.
• There is a need for novel, efficient, and accessible methods for TCM constitution identification.

Purpose of the Study:

• To establish a correlation model between Traditional Chinese Medicine (TCM) constitution and physical examination indexes using backpropagation neural network (BPNN) technology.
• To propose a new method for TCM constitution identification that addresses current limitations in clinical practice.

Main Methods:

• Data from 950 physical examinees, including physical examination indexes and TCM constitution types, were collected and classified.
• A backpropagation neural network (BPNN) algorithm was implemented using C# and Google's AI library to build and validate the correlation model.
• The dataset was divided into training and testing groups to evaluate the model's performance.

Main Results:

• The overall network model integrating physical examination indexes and TCM constitution achieved 88% accuracy in the training group and 53% in the test group, with an error of 0.001.
• Specific models for liver function, renal function, blood routine, and urine routine showed varying accuracies (31-60% for training, 38-42% for testing) and errors (2.4-11.7).
• Increased use of physical examination indexes in training improved the accuracy of the predictive network model.

Conclusions:

• A strong correlation exists between Traditional Chinese Medicine (TCM) constitution and physical examination indexes, as demonstrated by the developed model.
• The integration of physical examination indexes and TCM constitution through BPNN offers a novel approach for TCM constitution identification.
• This study provides a new pathway for TCM constitution identification, potentially overcoming existing challenges and promoting the integration of Chinese and Western medicine.