KR20110098286A - Self health diagnosis system of oriental medicine using fuzzy inference method - Google Patents

Abstract

The present invention is a self-diagnostic method of oriental medicine using fuzzy inference technique, when symptoms are found in the human body, the symptoms found can be easily inferred and diagnosed the disease at home early, By presenting the therapy, it is possible to treat the disease early.

Description

Oriental Self-diagnosis Method Using Fuzzy Inference Method {SELF HEALTH DIAGNOSIS SYSTEM OF ORIENTAL MEDICINE USING FUZZY INFERENCE METHOD}

The present invention relates to a method for self-diagnosis of oriental medicine using fuzzy inference technique, and more particularly, when symptoms are found in a human body, the symptoms found can be easily inferred and diagnosed at home in the disease. In addition, the present invention relates to a method for self-diagnosis of oriental medicine using fuzzy inference techniques that can treat diseases early by presenting folk remedies for the disease.

Traditionally, traditional Chinese medicine has been used to diagnose and heal Koreans' bodies. Unlike Western medicine, data on our bodies has accumulated over hundreds of years, and scientific medicine that can diagnose and prescribe diseases according to the unique physical characteristics of Koreans. to be. However, unlike traditional western medicine, oriental medicine is not widely used, so there is a sense of distance to the general public.

On the other hand, interest in one's health has recently increased. However, in order to use professional medical services due to busy living patterns and time and geographical conditions, psychological burden and time and economic costs are required, and it is difficult for workers and people living in the mountains to receive professional medical services.

In order to solve this problem, a self-diagnostic system using a neural network is conventionally used as shown in FIG. 1.

That is, the conventional self-diagnosis system using neural networks derives three final diseases by measuring the similarity between the symptoms selected by the user and the disease by applying the ART algorithm. In addition, after receiving the most prominent representative symptoms of the user, the symptoms of related diseases are provided as a query so that irrelevant symptoms are not selected.

However, if the symptoms of the disease is different from the early and late, or if the symptoms are many, such as a cold, but not all symptoms, there is a problem that the similarity between the disease and the symptom vector is lowered.

The present invention is to solve the above problems, the purpose is that when symptoms are found in the human body, the symptoms found can be easily inferred and diagnosed the disease at home early, the disease It is to provide a self-diagnosis method by using fuzzy inference technique that can treat diseases early by suggesting folk remedies.

Herbal self-diagnostic method using fuzzy inference technique according to the present invention for achieving the above object, comprising the steps of classifying the body parts into 17 parts to present to the user as a query; Classifying the symptoms using a fuzzy ART algorithm based on the symptoms selected by the user; Deriving the top five diseases with fuzzy inference rules; A step of presenting the derived top five diseases, the causes of the derived diseases, and the folk remedies presented in the oriental medicine books, wherein the processing of the fuzzy ART algorithm comprises: a first step of initializing each factor; Obtaining an output weight vector when the input vector is presented; Calculating a output vector and selecting a node having the largest output value as a winner node; A fourth step of verifying similarity by applying the output vector selected as the winner node and the boundary variable; Adjusting a learning rate and connection strength if similarity is recognized, and searching for new winner neurons and repeating the learning process if similarity is not recognized; If no cluster satisfies the similarity for all patterns, a sixth step of generating a new cluster is included.

According to the herbal self-diagnosis method using the fuzzy inference technique according to the present invention, when a symptom is found in a human body, the found symptoms can be easily inferred and diagnosed in the home at an early stage.

Other objects and advantages of the present invention in addition to the above object will be apparent from the detailed description of the embodiments with reference to the accompanying drawings.

1 is a view for explaining a self-diagnosis system using a conventional neural network.
Figure 2 is a process configuration diagram of a herbal self-diagnosis system using a fuzzy inference technique applied to the present invention.
3 is a view for explaining a data collection and database construction process applied to the present invention.
4 is a diagram showing an ER-Diagram of a database applied to the present invention.
5 is a flowchart of a fuzzy ART algorithm applied to the present invention.
FIG. 6 is a diagram illustrating a membership function for matching a user selected symptom and a disease symptom applied to the present invention. FIG.
7 is a diagram showing the membership function for the total number of symptoms included in the disease to be applied to the present invention.
8 is a view showing a function of belonging to a disease for a user-selected symptom applied to the present invention.
9 illustrates a fuzzy inference rule applied to the present invention.
10 is a symptom selection screen when a male applied to the present invention
11 is a symptom selection screen when a woman is applied to the present invention.
12 is a view showing a result screen for deriving a disease using a fuzzy ART algorithm applied to the present invention.
FIG. 13 is a diagram illustrating a result screen of a disease derived by applying a fuzzy inference technique to a fuzzy ART algorithm applied to the present invention. FIG.
14 is a view showing a process of creating a consultation request applied to the present invention.
15 is a view showing a process of creating a consultation result book applied to the present invention.

Hereinafter, with reference to the accompanying drawings will be described in detail the herbal self-diagnosis method using the fuzzy inference technique according to the present invention.

The process configuration of the herbal self-diagnosis system using the fuzzy inference technique applied to the present invention is shown in FIG.

In addition, the database applied to the present invention, after extracting the disease using the Korean Standard Disease Sign Classification Table (KCD), which replaced the International Disease Classification Table (ICD) defined by the World Health Organization (WHO) with the Korean Medicine Disease, Collect symptoms, causes and folk remedies for each disease through consent and other oriental medicine books.

The number of diseases extracted from the Korean Standard Disease Signage Classification Table (K.C.D) is 782 and the number of symptoms included in the extracted diseases is 414. The extracted diseases and symptoms are verified by a total of 3 oriental medicine specialists and the diseases and symptoms are built into a database.

The completed symptom is classified by body parts for presentation by the user's query. Data collection and database construction process is shown in FIG.

In addition, the database is composed of a disease table, symptom table, body parts table for learning and diagnosing diseases, member information table, specialist information table, consultation board table to manage the consultation service and member information of the Chinese self-diagnosis system It consists of a general bulletin board table. The disease table consists of ICD code, KCD code, Korean and Western disease names, disease cause, folk remedies, and symptom code for each disease. The ER-Diagram of the database is shown in FIG. 4.

On the other hand, in disease early learning using the fuzzy ART algorithm applied to the present invention, ART includes ART1 for clustering binary vectors and ART2 for clustering analog vectors.

ART1 cannot process analog vectors, and the performance of the system depends on the order of application of the input patterns. ART2, which is designed to classify analog vectors, has difficulty in selecting too many variables and has a high cost structure with a large amount of computation. Fuzzy ART generalized ART1 to handle analog vectors by replacing ART1's Product operation with the Fuzzy learning theory's min operation to solve these problems.

In the present invention, the processing of the fuzzy ART algorithm for application to oriental self diagnosis and learning is as follows.

Step 1: Initialize each argument.

Step 2: Given the input vector, obtain the output weight vector.

Step 3: Compute the output vector and select the node with the largest output value as the winner node.

Step 4: Verify the similarity by applying the output vector and boundary variables selected as winner nodes.

Step 5: If similarity is recognized, adjust learning rate and connection strength. If similarity is not recognized, search for new winner neurons and repeat the learning process.

Step 6: If no clusters satisfy the similarity for all patterns, create a new cluster.

In this case, the maximum output vector is set to 1 to solve a case in which the maximum value of the output vector does not become 1 because of an inconsistency between the input vector composed of binary vectors and the connection weight vector having continuous characteristics. As a weighting factor, it applies to the degree of belonging to the diagnosis of disease in the disease diagnosis and learning system.

The final output vector learned by applying the fuzzy ART algorithm means the degree of belonging of the disease to each symptom. The fuzzy ART algorithm flow chart is shown in FIG. 5.

Meanwhile, in the derivation of diseases using fuzzy inference techniques, the number of symptoms and the total number of symptoms of the diseases that match the symptoms selected by the user are applied to the fuzzy membership function to calculate the degree of belonging, and then the ratio is determined using the fuzzy inference rule. The degree of belonging of the disease is adjusted according to the symptom selected by the user by multiplying the fuzzy value by the output value of the first case.

In Equation (1), Output is the output value of each disease output by the fuzzy ART algorithm, and Ot is a value obtained by applying the total number of symptoms and the total number of symptoms corresponding to the symptoms selected by the user to the fuzzy inference technique. If the value of Ot is high, the value of Output is high because it is related to the symptom selected by the user. If the value of Ot is low, the value of Output is low because it is less related to the symptom selected by the user. The equation for obtaining Ot is shown in Equation (2).

The fuzzy affiliation interval for the number of symptoms of the disease that matches the user's selected symptoms is 5 sections. The number of symptoms of the disease that matches the symptoms selected by the user is small in the S section, ST in the slightly smaller section, M in the middle section, Segments that match a lot more are classified as BT and segments that match a lot are classified as B. The membership function for matching the user's selected symptom with the disease symptom is shown in FIG. 6.

In addition, the fuzzy affiliation section of the total number of symptoms included in the disease is a total of five sections, the S number of small symptoms included in the disease, ST is a little smaller, M in the middle, BT a lot, Classify as B The membership function is shown in FIG.

In addition, the output belonging function for calculating the Ot representing the degree of belonging of the disease to the user-selected symptoms is shown in FIG.

Meanwhile, a fuzzy inference rule for deriving a disease corresponding to a symptom selected by a user may be expressed in a fuzzy relationship in an if-then format as shown in FIG. 9.

The number of symptoms corresponding to the symptoms selected by the user and the total number of symptoms of each disease are applied to the fuzzy inference technique to obtain the degree of belonging of the diseases corresponding to the selected symptoms. The Max_Min inference method is applied to the degree of belonging of the obtained diseases. Finally, the degree of belonging of the inferred diseases is obfuscated to find the Ot value that finally derives the diseases.

On the other hand, the experimental environment of the herbal self-diagnosis system using the fuzzy inference technique applied to the present invention is Eclipse 3.2, Apache Tomcat 5.5, Apache 2.2, Adobe Potoshop 7.0, JDK on an IBM compatible PC equipped with Intel Pentium_IV 2GHz CPU and 1G RAM. 1.6, DBMS is implemented to use in On-line environment by JSP using Oracle 10g.

In addition, the proposed Chinese self-diagnosis system is implemented by classifying body image into 17 parts for the user to select the symptom. In addition, the body image according to the gender of the user is presented to eliminate the user's confusion. 10 is a symptom selection screen in the case of a male, and FIG. 11 is a symptom selection screen in the case of a female.

When the user selects fever, high fever, vomiting, cough, runny nose and nasal bleeding (nose bleeding) as a symptom using the existing fuzzy ART algorithm [9], Table 1 shows the result of analyzing the number of concordances among the symptoms corresponding to the derived diseases.

Disease name Number of matches whooping cough 5/15 Shingles 0/3 Anthrax 4/9 Queue 1/5 rabies 3/15

On the other hand, the result screen derived disease using the fuzzy ART algorithm is shown in FIG.

When the fever, high fever, vomiting, cough, runny nose and nasal bleeding (nose bleeding) were input as in the above method, the results of analysis of the five diseases derived by applying the fuzzy inference technique to the fuzzy ART algorithm are shown in Table 2 and same.

Disease name Number of matches Viral meningitis 3/3 Histoplasmosis 3/8 whooping cough 5/15 Anthrax 4/9 Yellow fever 3/11

Meanwhile, the result screen of applying the fuzzy inference technique to the fuzzy ART algorithm is shown in FIG.

As can be seen from the experiment and the analysis of the results, it can be seen that the results of applying the fuzzy inference method to the fuzzy ART algorithm to derive the disease have higher accuracy of disease derivation than the results of applying the existing fuzzy ART algorithm to the disease. .

The herbal self-diagnosis system applied to the present invention provides a counseling system through a 1: 1 counseling board between a general member and a professional member. The general member may request the opinion of a specialist by filling out a request for consultation in order to answer questions about the disease. The professional member can resolve the general member's curiosity by checking the consultation request written by the general member and completing the consultation result. The process of creating a consultation request form by a general member is illustrated in FIG. 14, and the process of creating a consultation result form by a professional member is illustrated in FIG. 15.

In the present invention, in order to improve the accuracy of disease deduction, the fuzzy inference technique is applied to the result of applying the number of symptom matches selected by the user for each disease and the total number of symptoms of each disease to the fuzzy ART algorithm to improve the disease derivation performance. It was.

In addition, in order to solve the general member's curiosity about the disease outbreak, a counseling system was provided through a 1: 1 counseling board between professional members so that the general member could solve the question about the result of the disease outbreak.

The database builds a disease database based on the Korean Standard Disease Signage Classification (KC.D), which replaces the International Disease Classification (ICD) created by the World Health Organization (WHO) with oriental medical diseases, providing users with increased credibility to the disease. I could do it. As a result of the actual analysis of the proposed oriental medicine self-diagnosis system, it is more likely to be applied to the self-diagnosis of the user than the conventional herbal self-diagnosis system. In addition, it was confirmed in the experiment and analysis of results that it can be used as a supplementary diagnostic tool of oriental medicine.

While the present invention has been described in connection with what is presently considered to be practical exemplary embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, Of course.

Claims (1)

Oriental Medicine Self-diagnosis Method Using Fuzzy Inference Method,
Classifying the body parts into 17 parts and presenting them to the user as a query; Classifying the symptoms using a fuzzy ART algorithm based on the symptoms selected by the user; Deriving the top five diseases with fuzzy inference rules; Including the top five diseases identified, their causes and the folk remedies suggested in the oriental medicine books,
The processing of the fuzzy ART algorithm may include a first step of initializing respective factors; Obtaining an output weight vector when the input vector is presented; Calculating a output vector and selecting a node having the largest output value as a winner node; A fourth step of verifying similarity by applying the output vector selected as the winner node and the boundary variable; Adjusting a learning rate and connection strength if similarity is recognized, and searching for new winner neurons and repeating the learning process if similarity is not recognized; If there is no cluster that satisfies the similarity for all the patterns, the self-diagnosis method using fuzzy inference method comprising the step of generating a new cluster.

Images