Research on correlation model between TCM Constitution and physical examination index based on BPNN algorithm

Authors names: 

Yue LUOa, Bing LINb*, Chuan-Biao WENa, Mao LUOa

Authors working units: 

a. Chengdu University of Traditional Chinese Medicine, Chengdu, Sichuan, 611137,China

b. The Affiliated Hospital of Chengdu University of Traditional Chinese Medicine, Chengdu, Sichuan,610075, China

Corresponding Author: 

Bing LIN

Corresponding Author Information: 

Bing LIN, director of Health Management Center of Teaching Hospital of Chengdu University of TCM, archiater, research oriented at health management; excel at TCM health management, constitution identification, and health maintenance & regulation guide, etc.

Email: 1284516264@qq.com

Abstract: 

This paper studies the correlation model between TCM Constitution discrimination and physical examination index based on BPNN algorithm.The use of BPNN algorithm, 253 cases of urine routine test, linkage model is constructed of TCM Constitution and physical indicators, according to the test, the correct rate of learning and test group are respectively 60% and 40%. It is proved that there is a strong correlation between TCM Constitution and physical examination indexes. With the help of modern science and technology to develop Chinese medicine, it can effectively improve the development level of TCM and promote the modernization process of TCM Treatment.

1. Introduction

As the understanding of Chinese medicine deepens worldwide, its ability to prevent and cure diseases has received increasing attention.  The role of traditional Chinese medicine has  recognized and support by the world health organization in the fight against all kinds of diseases[1]. Influenced by a mode of inheritance, the development and application of traditional Chinese medicine (TCM) has been relatively slow, and far from realizing its due value. Chinese Westernization, the loss of talented people, rarity, Houjifaren, the technical and price difference of TCMs, the imbalance of Chinese medical technology development, and limited medical staff are factors restricting the development of TCM. The speed in the development of TCM cannot meet the current demands of society for TCM.

    The Chinese Medicine Institute has issued a “Chinese constitution classification and determination standard.” According to this, many scholars have concluded that the clinical research of the current constitution can be divided into two aspects: one is to investigate the distribution of the constitution in disease, the other is to study the correlation between body mass and disease, which occurs mostly for the purpose of studying and preventing the development of the disease[2].

    The application of the back-propagation neural network (BPNN) in the field of the Chinese constitution is to fill the vacancy in the study of the TCM constitution identification field of neural networks; the use of information technology to promote the rapid development of Chinese medicine; and the development of a TCM constitution identification system, to serve the benefit of a wider population[3].

2. BPNN-based algorithm construction of traditional Chinese medicine constitution and physical examination index correlation

2.1 Characteristics of the constitution and physical examination indexes in traditional Chinese medicine

To ensure that the neural network input and output can accurately express the constitution and physical examination indicators, and to improve the convergence speed of the neural network, it is necessary to quantify the physical type and physical examination indicators of TCM[4].

2.1.1 Input information index of neural network

The examination type is the input of the neural network, with a digital representation of gentleness, Qi deficiency, Yang deficiency, Yin deficiency, phlegm, dampness, blood stasis, Qi stagnation, and special qualities to represent the nine types of the TCM constitution. Belonging to a physical type is denoted by 1, whereas 0 represents the constitution that the type does not belong to. Therefore, there are nine different TCM constitutions with digital representation as shown in Table 1.

 

Table 1. Digital representation of traditional Chinese medicine(TCM) constitution types

Order number

Somatotypes

Digital representation

1

Peace quality

100000000

2

Qi deficiency constitution

010000000

3

Yang deficiency

001000000

4

 

Yin Asthenia

000100000

5

Phlegm dampness

000010000

6

damp-heat constitution

000001000

7

Blood stasis syndrome

000000100

8

Qi stagnation constitution

000000010

9

Special quality

000000001

 

2.1.2 Neural network output information index

According to the conventional physical examination, the examination indexes are divided into: basic information (age, sex, patient height, weight, body mass index, systolic blood pressure, diastolic blood pressure), blood routine index (white blood cell count, neutrophil count, lymphocytes, monocytes, eosinophils, basophils, the percentage of neutrophils and lymphocytes, mononuclear cells, the ratio of the percentage of eosinophils, basophil, erythrocyte, hemoglobin, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, red cell distribution width coefficient of variation, red cell distribution width, the degree of liver function, platelets, total protein, albumin, globulin, albumin/globulin, alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, glutamyl transferase, total bilirubin, direct bilirubin, and indirect bilirubin), renal index (urea nitrogen, creatinine, uric acid, glucose, and carbon dioxide combining power), blood lipids (cholesterol, triglycerides, high density lipoprotein cholesterol, and low density lipoprotein cholesterol).

    The quantization index of each output parameter is represented by Ai, and the measured value is expressed as Mi. All the output quantization indexes are limited to the range of [0,1], and the quantitative expression of the basic information of the patient is as follows:

\(A_{age}=M_{age}/100, A_{sex}=0(man)|1(woman), A_{height}=M{height}/1000\\ A_{weight}=M_{weight}/100, A_{BMI}=M_{BMI}/100, A_{Systolic Blood Pressure}=M_{Systolic Blood Pressure}/1000\\ A_{Diastolic Blood Pressure}=M_{Diastolic Blood Pressure}/100\)

2.2 Brief introduction of BPNN

Neural cell neural network simulation is similar to the way that a human brain works but greatly simplified. The network is composed of many neural network layers, and each layer is composed of many elements. The first layer is the input layer, the last layer is called the output layer, each layer is called a hidden layer in the BPNN. There is a link between each unit only in the adjacent nerve layer. In addition to the output layer, each layer has a bias node.

2.2.1 BPNN overview

BPNN is a multilayer feedforward network,and it’s also known as back-propagation neural network,it has a highly nonlinear mapping relationship between the input and output of the recognition model. To achieve a nonlinear classification and approximate arbitrary nonlinear function with arbitrary precision,we can transform the activation function or modify the threshold and the number of the layers of the BP neural network. BPNN can learn through samples, and automatically adjust the weights in the network, so as to realize the non-inductive logic.

2.2.2 The principle of BPNN

BPNN is one of the most widely used neural network models. The BPNN can learn and store a lot of input-output model mapping to reveal the mathematical equations describing the mapping relation. Its learning rule is the steepest descent method used to adjust the weights and thresholds of the network through the BPNN, the minimum error sum of squares.

2.2.3 BPNN advantage analysis

The BPNN maps from input to output function, the existing mathematical theory has proved that it can achieve any nonlinear mapping ability for complex nonlinear problems, the neural network can be regarded as a "black box", to facilitate the internal mechanisms[5]. Moreover, the subsystems are relatively independent and need no decoupling. It has a very strong fault-tolerant ability and self-learning ability.

 

2.2.4 BPNN algorithm

In the \(ij\) hidden layer input vector, \(jk\)is the hidden layer output vector, \(w_{ij}\)is the input layer and the middle layer connection weights, \(\theta_j\)is the hidden layer neurons and $$\sum_{i=1}^n w_{ij}x_{i}$$ is the hidden layer threshold, the input vector of  nerve cell \(i\) is \(^{x_i}(i=1,2,...,n)\).

$$h_j=\sum_{i=1}^nw_{ij}x_i-\theta_j=\sum_{i=1}^{n+}w_{ij}x_i$$

    The output function of hidden layer uses the \(f(h_j)\) excitation function

\(o_j=f(h_j)=1/(1+e^{-h_j})\)

    In the formula:\(j=1,2,...,m;\) \(\theta_j=w(n+1)jxn+1, xn+1=-1\)\(y_k\)is the threshold of each neuron in the output layer. The input of the output layer K node is \(k\), and the output layer uses \(f(h_k)\)excitation function:

\(h_k=k\\ y_k=f(h_k)=1/(1+e^{-h_k})\)

    In the formula:\(k=1,2\), the BP algorithm is a supervised learning algorithm, input the learning sample data for p \((x1,x2,...,x3)\), and the corresponding monitoring object is \(t1,t2,...,t_p\), the learning algorithm will be the actual output \(y1,y2,...,y_p\)and \(t1,t2,...,t_p\) error to modify its weight and threshold, so that \(y_p\) and \(t_p\) are as close as possible. The \(w_{jk}\) is the connection weight value of the implicit layer and the input layer, and \(\eta\) is the step size. Set \(n_0=5000\) for the number of iterations, using the output of the \(\sigma_{jk}^{p_i}\)  and hidden layer neurons of each neuron of the output layer to correct the connection weight, and the correction formula is as follows:

$$w_{jk}^{n_0+1}=w_{jk}^{n_0}+\eta\sum_{p_1=1}^p\sigma_{jk}^{p_1}o_j^{p_1}$$

    In the formula: \(y_p\) is the actual output value, \(t_p\) is the error value, \(o_j\) is the threshold of each neuron of the hidden layer, and uses the output \(\sigma_{jk}^{p_i}\) neurons of the hidden layer and input layer of each neuron to correct the connection weight, the correction formula is as follows:

$$\sigma_{p_1}^{jk}=(t_{p_1}^k-y_{p_1}^k)y_{p_1}^k(1-y_{p_1}^k)\\ \sigma_{ij}^{p_1}=\sum_{k=1}^k\sigma_{jk}^{p_1}w_{jk}o_j^{p_1}(1-o_j^{p_1})$$

    In the formula, \(t_k\) is the desired output vector and \(y_k\) is the output layer output vector. The BP algorithm uses a gradient descent algorithm to update the weights in the network. If a batch update algorithm is used, the batch size is set to \(p\), a square error and calculation formula are used, and then the global error of the batch is:

$$E=\frac{1}{2}\sum_{p_1}^p\sum_{k=1}^k(t_k^{p_1}-y_k^{p_1})^2<\varepsilon$$

3. Testing and verification

3.1 The process of the neural network algorithm for comparing the constitution and physical examination index of traditional Chinese medicine

    According to the demand of the correlation model between the TCM constitution and physical examination index, and the characteristics of BPNN, the algorithm flow of the graph network model is established.

    The constitution of the TCM and Medical Association index of the learning algorithm of NN process is as follows[6]

    1) The structure of BPNN is constructed. The network model includes an input layer (physical examination index), hidden layer and output layer (physical type).

    2) The weights and thresholds of the BPNN are initialized, and the steps and accuracy of learning are determined.

    3) A group of Chinese medicine and physical indicators of learning samples (such as the constitution TCM and blood lipid index) is entered for each sample (input vector and the desired output) for learning.

    4) Connection weights and threshold input vector calculates the value of activation of neurons in the hidden layer. The hidden layer activation function calculates the output value of each unit.

    5) Connect weights and thresholds with Zi, calculate the activation of output layer , then use activation function, calculate the output Yjof output layer.

    6) The calculation of expected output error, hidden layer output and the actual output from the output layer, BP, modify the weights and thresholds layer by layer.

    7) Learning the sample data until the sample data is finished, and judging whether the global error is within the given range of accuracy. This is to complete the training of the sample data, not to test the next set of data.

    8) Save the threshold,complete the construction of the model.

3.1.1 Collect system input data

The system input data mainly includes blood, blood lipid, liver function, and kidney function of the four modules of the data acquisition. The study included 500 subjects from the hospital of traditional Chinese medicine. The TCM constitution types and examination indicators were obtained to assess the corresponding relationship between the samples. With basic personal information hidden, the data entry staff input data into an Excel table, then the data were entered into an Access database.

3.1.2 Quantitative result data

The BPNN model test output value is nine types of physique, through testing multiple sets of data and recording the test results to determine the accuracy.

3.2 Test of physique type

3.2.1 Test

The blood urea nitrogen, creatinine, uric acid, glucose, carbon dioxide binding force of several indicators was used to test the physique, the sample data of 253 groups were selected to determine the type of constitution.

3.2.2 Result analysis and documentation

In diagnosis on the human body using TCM the results are often unlike Western medicine, but in general, there is not a particularly large difference. Using the algorithm, the determination of the results was 60%, this needs to be optimized to improve the correct rate. After the test, the results are recorded.

3.3 Improvement of the BPNN algorithm

Through the blood lipid data detection, the current test results were 60%, which is not that high. There is a need to continually improve and perfect the code to obtain a higher accuracy.

4. Discussion

According to the existing constitution, the identification, application and development status quo and problems in the process of Chinese informatization, the BPNN algorithm was used to construct a constitution identification and examination index model. After the system construction, we collected historical data to train, optimize and collect data to test and analyze the results. It is hoped that the application of this system can improve the scope of the TCM constitution identification service and application, and improve the level of TCM informatization construction.

Funding Support

This work was supported by Youth Special Fund for research on Traditional Chinese Medicine Science and Technology in Sichuan (Grant No.2016Q065).

Competing Interests

The authors declare that they have no conflict of interest.

Ackknowledgement

None.

References

  • 1. Bai Lina.Study on TCM Constitution identification based on BP neural network[D].Tianjin University of Technology.2014.2:1-2,27-28.
  • 2. Zhang Sida and Chen Songbai. Discussion on the essence and clinical significance of TCM Constitution identification. Shanxi traditional Chinese medicine 2012.28 (3):2.
  • 3. Yu Hao. Information technology helps the development of Chinese medicine. Chinese Medical Journal.2006.11.4:1.
  • 4.

    Yang Jing, Xing Tong and Li Chunlu. Research progress on modernization of TCM constitution theory [J]. Journal of Changchun Academy of traditional Chinese medicine. 2000.16:1-2.

  • 5. Guo Leilei. Signal detection and analysis based on neural network[D]. Wuhan University of Technology.2001.5.1:6-10.
  • 6. Blog Park.BP neural network and learning algorithm.2012.06.05:1.
Figures: 
中文摘要: 

本文研究基于BPNN算法的中医体质辨证与体检指标的关联模型。运用BPNN算法分析253例尿常规指标,构建中医体质与体检指标关联模型,经过测试验证,学习组和测试组的正确率分别为60%40%。证明中医体质与体检指标之间有较强关联性,为中医治未病的现代化进程提供参考。