CN112786197A - Traditional Chinese medicine pathogenesis network construction method and system based on network syndrome differentiation - Google Patents

Traditional Chinese medicine pathogenesis network construction method and system based on network syndrome differentiation Download PDF

Info

Publication number
CN112786197A
CN112786197A CN202110149996.2A CN202110149996A CN112786197A CN 112786197 A CN112786197 A CN 112786197A CN 202110149996 A CN202110149996 A CN 202110149996A CN 112786197 A CN112786197 A CN 112786197A
Authority
CN
China
Prior art keywords
network
pathogenesis
disease
chinese medicine
traditional chinese
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110149996.2A
Other languages
Chinese (zh)
Other versions
CN112786197B (en
Inventor
许强
温川飙
李炜弘
孙涛
赵姝婷
张松
高原
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chengdu University of Traditional Chinese Medicine
Original Assignee
Chengdu University of Traditional Chinese Medicine
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Chengdu University of Traditional Chinese Medicine filed Critical Chengdu University of Traditional Chinese Medicine
Priority to CN202110149996.2A priority Critical patent/CN112786197B/en
Publication of CN112786197A publication Critical patent/CN112786197A/en
Application granted granted Critical
Publication of CN112786197B publication Critical patent/CN112786197B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/04—Inference or reasoning models
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Abstract

The invention discloses a traditional Chinese medicine pathogenesis network construction method based on network syndrome differentiation, which is a digital syndrome differentiation method taking pathogenesis as a core and system science as guidance. The network syndrome differentiation generates all candidate pathogenesis networks based on the concept of the systematic science reduction theory; screening out a limited integral pathogenesis network based on the overall concept of system science; and determining the unique clinical pathogenesis network by combining a man-machine cooperation strategy. The fitting and restoration of scientific connotations such as traditional Chinese medicine system view, overall view, individuation and the like can be realized through the networked definition of traditional Chinese medicine pathogenesis, so that the scene applications such as traditional Chinese medicine digital inheritance, digital diagnosis and treatment and the like can be better realized. The utility model provides a disease mechanism network traditional chinese medicine differentiation system based on network differentiation, can be through the input symptom, output disease mechanism network realizes the comprehensive, expression directly perceived to the disease mechanism, greatly assists the excavation of clinical differentiation rule, and rethread man-machine cooperation module reachs individualized disease mechanism network, has improved the reliability of differentiation result.

Description

Traditional Chinese medicine pathogenesis network construction method and system based on network syndrome differentiation
Technical Field
The invention relates to the field of digital syndrome differentiation of traditional Chinese medicine, in particular to a traditional Chinese medicine pathogenesis network construction method and system based on network syndrome differentiation.
Background
With the interpenetration and crossing of modern biology, engineering discipline and informatics, the research of biological problems by means of thinking of "system" and "integration" driven by system biology is pushed to the frontier of life sciences. The traditional Chinese medicine considers the human body as a complex huge system, and the Qian schensen indicates that the human science has a systematic view for many times, which is the view of the traditional Chinese medicine. Wang Yongyan academicians and Daidayi academicians also emphasize that the research of traditional Chinese medicine should be based on the scientific thought of the system. Modern old and young people, who congratulate on the disease and the turner, adopt the theory and method of modern systematic science to dig up the theory of systematic theory in the theory and practice of traditional Chinese medicine, and develop the important research field of systematic traditional Chinese medicine. However, the systematic science itself has not made a breakthrough development because of the development difficulty of the systematic science itself. However, it is undeniable that the general understanding of traditional Chinese medicine by Qiao Lao is accurate, and the theory of the system is the core scientific connotation inherent but neglected in traditional Chinese medicine.
The traditional Chinese medicine considers the human body as an organic whole formed by taking the five internal organs as the center and matching with six internal organs through the communication function of the meridian system. Physiologically, TCM considers the human body as a functional network of five zang-organs and six fu-organs, and the systems are restricted and coordinated with each other to maintain physiological balance. In the pathological aspect, the disease or the mechanism of occurrence and development is considered as the result of the combined action of qi and deficiency of vital qi and (or) pathogenic qi, and the qi and deficiency of vital qi can be subdivided into the pathogenesis of lung-yin deficiency and kidney-qi deficiency due to the difference of viscera, yin-yang and qi-blood; because of the different types of pathogenic factors, the exuberant pathogenic factors can be subdivided into wind-cold exogenous pathogenic factors and phlegm-turbid pathogenic factors. Because there are extensive binary causal relationships among the pathogenesis elements, i.e., the disease chain, such as deficiency of heart qi due to deficiency of lung qi and internal dampness due to deficiency of spleen qi, it is obvious that the essence of the pathogenesis is a complex network formed by the pathogenesis elements and the disease chain.
From the perspective of the system theory, the concept of the syndrome of traditional Chinese medicine is questionable, and there is no concept of the syndrome in the historical development process of traditional Chinese medicine, and the syndrome is a concept appeared in the last 50 th century. Syndrome differentiation of traditional Chinese medicine is influenced by the diagnosis mode of western medicine diseases, which facilitates the propagation of traditional Chinese medicine, but violates the basic scientific principle of traditional Chinese medicine. Taking fig. 1 as an example, fig. 1 shows a series of pathological changes of deficiency of lung qi, wind-cold affection from the exterior, wind-cold entering the interior to transform heat, and deficiency of lung qi resulting in deficiency of heart qi, kidney qi, spleen qi and the like. Obviously, the conventional four-character syndrome type structure is difficult to comprehensively express the complex pathogenesis, even if the complex pathogenesis can be forcibly expressed by 'heat phlegm turbidity in wind cold due to deficiency of heart, lung, kidney and spleen qi', the structure also loses the relation information between pathogenesis elements, and for the relation information of 'spleen yang deficiency causing kidney yang deficiency or kidney yang deficiency causing spleen yang deficiency', the spleen kidney yang deficiency 'is lack of the relation information for expressing the spleen yang deficiency', and the relation information is the important composition content of the pathogenesis concept. No matter scientific research or clinical treatment, the traditional Chinese medicine diagnosis and treatment modes of ' one disease and several symptoms, and ' several prescriptions ' need to be considered again.
With the development of informatics, combining traditional Chinese medicine with modern computing technology has always been an important way for Chinese medicine to seek breakthrough. TCM emphasizes syndrome differentiation, while digital syndrome differentiation is the key and prerequisite of informatization. In the field of digital syndrome differentiation, syndrome differentiation is an important research system, the core concept of syndrome differentiation is 'dimension reduction and order increase', clinically any complex pathogenesis can be 'dimension reduction' decomposed into the most basic diagnosis units (etiology, disease position and disease property), and the most basic diagnosis units can be mutually combined to form the complex pathogenesis. Based on the thought, syndrome differentiation provides a three-order double-network syndrome differentiation medical model of 'symptom-syndrome type', namely, the symptom determines the syndrome through the relation between the sum of the symptom-syndrome weight and the threshold value, and the syndrome is combined to form the final syndrome type. However, the essence of the system theory in traditional Chinese medicine is the relationship between the study objects, and the simple decomposition and combination of the syndrome elements separate the relationship between the syndrome elements, for example, "spleen deficiency with excessive dampness" and "dampness with excessive spleen deficiency" in the theoretical system of traditional Chinese medicine are two different pathogenesis, but they are expressed as the same syndrome element combination form (spleen, dampness, qi deficiency) in the syndrome differentiation system. For example, the core step of syndrome differentiation- "determining the syndrome through the relationship between the sum of symptom-syndrome weight and threshold" can be expressed mathematically as a function y ═ Wx + b, where the matrix W represents all symptom-syndrome weights, b represents the threshold, the vector x represents all input symptoms, and the vector y represents the discriminant state (presence or absence) of the syndrome, and from an informatics perspective, Wx + b is a linear function, but the chinese medicine itself is nonlinear.
In recent years, with the development of artificial intelligence techniques such as machine (deep) learning, many scholars try to introduce them into the field of digital syndrome differentiation of traditional Chinese medicine, however, most of the studies are based on the syndrome differentiation mode of "syndrome type" classification, and as mentioned above, the syndrome itself is questionable. From the aspect of informatics, almost all machine learning techniques solve the classification problem (regression, clustering and other problems can be regarded as classification problems in nature), however, since the class (pathogenesis) and class (pathogenesis) of traditional Chinese medicine have a wide network relationship, the pathogenesis has no class concept due to the network relationship to a certain extent. In addition, the loss function is an inevitable element of all machine learning technologies, the essence of the loss function is the error between the prediction output of the metric model and the real standard output, however, due to the individualized characteristics of the traditional Chinese medicine, the same set of symptom samples are often the real output without the standard, and the error and the loss cannot be defined.
In summary, the syndrome differentiation or the digital syndrome differentiation based on the machine (depth) learning technique has the problem of insufficient fitting to the medical meaning, which limits the application of the method in the real clinic.
Disclosure of Invention
The invention aims to: aiming at the problem that the prior art has insufficient fitting to the connotation of Chinese medicine in the syndrome differentiation mode of syndrome differentiation or the digital syndrome differentiation mode based on the machine (depth) learning technology, the traditional Chinese medicine pathogenesis network construction method and system based on network syndrome differentiation are provided.
In order to achieve the purpose, the invention adopts the technical scheme that:
a traditional Chinese medicine pathogenesis network construction method based on network syndrome differentiation comprises the following steps:
s1, generating a set PNs of all candidate pathogenesis networks according to all the symptom-pathogenesis binary relations and pathogenesis-pathogenesis binary relations;
s2, according to the accessibility of one pathogenesis element in the candidate pathogenesis network to all other pathogenesis elements, the overall property of the pathogenesis network is judged, the same judgment operation is carried out on all candidate pathogenesis networks, and the limited overall pathogenesis network set h-PNs is screened out from the set PNs.
A traditional Chinese medicine pathogenesis network construction method based on network syndrome differentiation is a digital syndrome differentiation method which takes pathogenesis as a core and system science as guidance. The network syndrome differentiation generates all candidate pathopoiesia networks based on the theory of systematic science reduction, screens out limited integral pathopoiesia networks based on the overall concept of systematic science, can jump out a traditional Chinese medicine syndrome type syndrome differentiation mode through the networked definition of traditional Chinese medicine pathogenesis, realizes the fitting and reduction of traditional Chinese medicine system view and overall concept of the science, and better realizes the scene application of traditional Chinese medicine digital inheritance, digital diagnosis and treatment and the like.
Preferably, in step S1, all the binary relationships between symptom and pathogenesis and the binary relationships between pathogenesis and pathogenesis are stored in the matrix C and the matrix M, respectively, a set of clinical random symptom groups S is given, and the algorithm for generating all candidate pathogenesis network sets PNs corresponding to S specifically includes:
s11 reading symptoms S from matrix C respectivelyiCorresponding candidate pathogenesis elements are respectively stored in the set CiIn respect of all CiExecuting Cartesian product operation, and exhaustively enumerating to obtain all disease pathogenesis element sets V, i.e. generating
Figure BDA0002931930360000041
Set of candidate disease pathogenesis, symptom SiHas | CiSelecting | species;
s12 reading V from matrix MiAll of correspondingDisease chain<vm，vn>And storing the disease mechanism chains to a disease mechanism chain set Ei(ii) a Wherein, the pathogenesis chain<vm，vn>Is a binary causal relationship between disease mechanism elements;
s13 integrating disease pathogenesis ViThe chain of pathogenesis EiCoupling to obtain all candidate pathogenesis networks, wherein the pathogenesis elements are collected ViSet of vertices, chains of pathogenesis E, representing the networkiRepresenting the edge of the network, and storing the edge into a set PNs to obtain the edge
Figure BDA0002931930360000042
And (5) planting candidate disease machine networks.
Preferably, in step S2, the determining operation performed on any candidate disease-causing network includes:
starting to judge the accessibility of the disease element to all other disease elements from the first disease element of the network, if the disease elements are all accessible, marking the disease elements as base points, marking the network as an integral disease network, adding the network to the set h-PNs and terminating the cycle; if not, the above operations are repeatedly executed to judge the next pathogenesis element, and if the last vertex is not the base point, the network is judged to be a non-integral pathogenesis network.
Preferably, the step S2 further includes: for each candidate pathogenesis network GiAnd judging according to the number of pathogenesis elements with the degree of incidence of 0:
if it is
Figure BDA0002931930360000043
Then G isiA non-global pathogenesis network;
if it is
Figure BDA0002931930360000044
V is judged by Dijkstra algorithmrAccessibility to other pathogenesis A (v)r) If A (v)r) Is true, then vrIs a base point, GiTo the global pathogenesis network, GiIs added to G+；
If it is
Figure BDA0002931930360000051
There is a directed ring G in the networksFrom GsGet the pathogenesis v of the conception vesselrV is judged by Dijkstra algorithmrAccessibility to other pathogenesis A (v)r) If A (v)r) Is true, then vrIs a base point, GiTo the global pathogenesis network, GiIs added to G+。
Preferably, the method further comprises the following steps:
s3 adopts man-machine cooperation strategy, network syndrome differentiation determines the only individual disease network from the limited overall disease network.
A traditional Chinese medicine syndrome differentiation system based on a network syndrome differentiation pathogenesis network comprises at least one processor and a memory which is in communication connection with the at least one processor; the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of the above.
The utility model provides a disease mechanism network traditional chinese medicine differentiation system based on network differentiation, can be through the input symptom, output disease mechanism network realizes the comprehensive, expression directly perceived to the disease mechanism, greatly assists the excavation of clinical differentiation rule, and rethread man-machine cooperation module reachs individualized disease mechanism network, has improved the reliability of differentiation result.
Preferably, the system further comprises a human-computer interaction module.
In summary, due to the adoption of the technical scheme, the invention has the beneficial effects that:
the invention discloses a traditional Chinese medicine pathogenesis network construction method based on network syndrome differentiation, which is a digital syndrome differentiation method taking pathogenesis as a core and system science as guidance. The network syndrome differentiation generates all candidate pathogenesis networks based on the concept of the systematic science reduction theory; screening out a limited integral pathogenesis network based on the overall concept of system science; and determining the unique clinical pathogenesis network by combining a man-machine cooperation strategy. A traditional Chinese medicine pathogenesis network construction method based on network syndrome differentiation can jump out a traditional Chinese medicine syndrome type syndrome differentiation mode through network definition of traditional Chinese medicine pathogenesis, and realize fitting and restoration of scientific connotations such as traditional Chinese medicine system view, overall view, individuation and the like, so as to better realize scene applications such as traditional Chinese medicine digital inheritance, digital diagnosis and treatment and the like.
The utility model provides a disease mechanism network traditional chinese medicine differentiation system based on network differentiation, can be through the input symptom, output disease mechanism network realizes the comprehensive, expression directly perceived to the disease mechanism, greatly assists the excavation of clinical differentiation rule, and rethread man-machine cooperation module reachs individualized disease mechanism network, has improved the reliability of differentiation result.
Drawings
FIG. 1 is a schematic diagram of the global pathogenesis network of the present embodiment.
FIG. 2 is a graphical illustration of a patient network.
FIG. 3 is a flow chart of the method of the present invention.
Fig. 4 is a schematic diagram of ALGORITHM 1.
Fig. 5 is a schematic diagram of ALGORITHM 2.
Fig. 6 is a schematic diagram of ALGORITHM 3.
Fig. 7 is a schematic structural diagram of the system provided in this embodiment.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings.
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Example 1
Definition of one, sick network
Traditional Chinese medicine knows life and diseases integrally from the perspective of systematic theory, and Qian schensen emphasizes that 'human science must have systematic view, which is the viewpoint of traditional Chinese medicine'. Structurally, the traditional Chinese medicine considers the human body as an organic system formed by taking the five zang organs as the center and matching with six fu organs through the communication of the meridians and collaterals. Physiologically, TCM considers the human body as a functional network of five zang-organs and six fu-organs, and the systems are restricted and coordinated to maintain physiological balance. Pathologically, the traditional Chinese medicine considers that the occurrence and development mechanism (i.e. pathogenesis) of the disease is the result of the interaction between pathogenic qi and healthy qi. From the viewpoint of systematic science, the pathogenesis can be defined as the set of basic pathogenesis (collectively called pathogenesis) corresponding to the deficiency of pathogenic factors and/or vital qi.
In fact, qi deficiency of the vital qi can be subdivided into the pathogenesis of kidney qi deficiency, kidney yin deficiency, heart blood deficiency and the like due to the difference of yin and yang, qi and blood, viscera and the like; due to different kinds of pathogenic factors, the exuberant pathogenic qi can be further divided into wind-heat exogenous pathogenic factors, damp pathogen and blood stasis. Under the theory system of traditional Chinese medicine, there are extensive binary causal (or subordinate) relationships between disease mechanism elements, for example, lung qi deficiency leads to spleen qi deficiency, and lung heat depends on interior heat, and we define the binary relationship between disease mechanism elements as the disease mechanism chain. The pathogenesis of the disease chain is further mathematically characterized by a complex network consisting of pathogenesis elements and a pathogenesis chain due to the transmission of the relationship between the pathogenic factors of the lung and the pathogenic factors of the wind-cold. The topic group defines the pathogenesis concept of traditional Chinese medicine as a pathogenesis network mathematically through the directed graph structure of graph theory, wherein the pathogenesis elements represent the top points of the network, and the pathogenesis chains represent the edges of the network. FIG. 1 illustrates a disease mechanism network consisting of 7 disease mechanism elements and 7 disease mechanism chains. The pathogenesis network expresses pathogenesis more comprehensively and intuitively, and most importantly, the expression of complex pathogenesis is realized through the networked definition of pathogenesis.
Secondly, introducing an integral pathogenesis network according to an integral reasoning principle
The research object of the system science is a complex system formed by a group of connected subsystems, the system science considers the connected subsystems as a whole, and the reduction theory and the whole are regarded as the basic method of the system science, and the method is highly consistent with the whole view idea of the traditional Chinese medicine. Under the guidance of system science, this section will discuss the application of the overall concept in syndrome differentiation of traditional Chinese medicine as an example.
A simple case was introduced, assuming that the practitioner of traditional chinese medicine collected two symptoms-superficial pulse and headache, while assuming that the practitioner had two prior knowledge (without considering the completeness of the prior knowledge at all): superficial pulse may be caused by wind-cold exogenous pathogenic factor (A) or wind-heat exogenous pathogenic factor (B); headache may be caused by wind-cold exogenous pathogens (A) or deficiency of kidney essence (C). Obviously, based on the principle of univalent reasoning, the practitioner can reason the wind-cold exogenous pathogenic factor (A) as the most probable cause of the case. The univalent reasoning principle in this case means: one symptom maps multiple candidate causes (pathologists), and the human brain always tends to reason the unique and same causes (pathologists) when multiple symptoms coexist.
Complicating the above cases, it is assumed that in addition to superficial pulse and headache, there is a third symptom of red tongue, while the practitioner has a third priori knowledge: a red tongue can be attributed to interior heat (D) or yin deficiency (E). If the proper syndrome differentiation result cannot be obtained according to the principle of univalent reasoning, but if the practitioner has other prior knowledge: pathogenic wind-cold can be transmitted into the interior to transform into heat, i.e. the disease chain (pathogenic wind-cold)
Figure BDA0002931930360000081
Interior heat). In this case, based on the overall reasoning principle, the practitioner can reason (wind-cold exogenous pathogenic factor (A) + interior fever (D)) as the most probable pathogenesis of the case. The overall reasoning principle in the present case means: one symptom maps multiple candidate causes (pathogenesis), and when multiple symptoms coexist, the human brain always tends to reason out a group of causes (pathogenesis) with overall connection. According to the integral reasoning principle, an integral pathogenesis network is further obtained from the pathogenesis network, and the method has high universality and is highly consistent with the integral concept of the traditional Chinese medicine.
Third, network syndrome differentiation
The invention provides a new digital syndrome differentiation method taking pathogenesis as a core, namely network syndrome differentiation, based on the overall concept and combined with the theory of graph theory. The network syndrome differentiation can be expressed as formula 1, wherein the matrix C represents the prior knowledge 1, i.e. the relationship between symptoms and pathogenesis elements (S2P for short), the matrix M represents the prior knowledge 2, i.e. the relationship between pathogenesis elements and pathogenesis elements (P2P for short), x represents the input symptom set, y represents the pathogenesis network, f1，f2，f 33 inference functions are represented. Need toIt is emphasized that the prior knowledge, i.e., S2P and P2P, is the premise of network syndrome differentiation, and the above relationship information is determined by voting by a plurality of doctors with the professional background of traditional Chinese medicine (delphi' S principle).
y＝f3(f2(f1(x：C，M)) (1)
The medical model of network syndrome differentiation comprises three stages: s1, exhaustively enumerating all candidate pathopoiesia networks based on Cartesian product operation, and S2, screening out limited overall pathopoiesia networks based on the overall concept of the traditional Chinese medicine; and the third part determines a unique individualized patient network based on the man-machine cooperation strategy. Step S2 i.e. f2Is the key to network syndrome differentiation. As mentioned above, in the process of syndrome differentiation in traditional chinese medicine, an effective pathogenesis network is screened out based on the holistic view principle, and in our research, we have established a heuristic mathematical definition (i.e. definition 1) for the holistic view principle. The overall view definition is the core of network syndrome differentiation. Theoretically, determining the integrity of the network is equivalent to determining whether (at least one) base point is present in the network. FIG. 2 is a graphical illustration of a pathogenesis network, we can observe the existence of a base point (P1) in G1; a base point (P1) exists in G2; a base point (P1) exists in G4; a base point (P1, P2, P6) exists in G5; g3, Gi have no base points. By definition we can draw the conclusion directly: g1, G2, G4 and G5 are integrated pathogenesis networks.
Here, definition 1 is proposed: if there is at least one vertex v in the pathogenesis network G that is reachable to other vertices, we specify that G has integrity and refer to it as the integrity pathogenesis network; otherwise G is referred to as non-global pathogenesis network. If a vertex v exists, we call it the base point of the network.
Then, a method for constructing a traditional Chinese medicine pathogenesis network based on network syndrome differentiation specifically comprises the following steps:
s1, generating a candidate pathogenesis network;
given a set of clinically random symptoms S, assuming that all symptom-pathogenesis binary relationships and pathogenesis-pathogenesis binary relationships are known and stored in the matrix C and the matrix M, respectively, the goal at this stage is to generate all candidate disease network sets PNs corresponding to S. The corresponding ALGORITHM is presented in ALGORITHM 1 in the form of pseudo code, as shown in fig. 4:
first, the symptoms S are read from the matrix CiCorresponding candidate pathogenesis elements are respectively stored in the set Ci(Line 1 in fig. 4). Then for all CiPerforming Cartesian product operation, and exhaustively enumerating to obtain all disease element sets V (Line 3-5 in FIG. 4), wherein the symbols
Figure BDA0002931930360000093
Representing a cartesian product operation. Is generated by this step
Figure BDA0002931930360000091
This candidate set of disease elements is due to the first symptom having | C1The candidate pathogenesis can be selected, and the second symptom corresponds to | C2Symptom S is selectediHas | CiAnd | selecting. V is then read from the matrix MiAll corresponding disease mechanism chains<vm，vn>And storing the disease mechanism chains to a disease mechanism chain set Ei(Line 7 in FIG. 4). Then, the pathogenesis elements are collected ViThe chain of pathogenesis EiCoupled, we obtain all candidate pathogenesis networks, wherein the pathogenesis elements are collected ViSet of vertices, chains of pathogenesis E, representing the networkiRepresenting the edge of the network (Line 8 in fig. 4). Finally, the same operation is carried out on all the disease element sets, all the candidate disease networks are obtained and stored in the set PNs (Line 9 in FIG. 4), and the total exhaustive enumeration is obtained through the stage
Figure BDA0002931930360000092
And (5) planting candidate disease machine networks.
S2 screening of integral pathogenesis network
Step S1 obtains all candidate pathologist network sets PNs, and step S2 aims to screen out the overall pathologist network h-PNs from the sets PNs. The pseudo code of the ALGORITHM is shown in ALGORITHM 2, FIG. 5, wherein A (v)m) Representing pathologists v in a networkmReachability to all other disease-causing elements in the network，A(vm) Is of the Boolean type, wherein "1" indicates all reachable, "0" indicates not all reachable, and the pathogenesis v is agreedmAnd is reachable by itself. According to definition 1, the screening principle is based on the holistic principle, and determining the integrity of the network is equivalent to determining whether a base point exists in the network. Specifically, given a sick network, we judge the reachability from the first sick element of the network to all other sick elements (Line 3 in fig. 5), if all sick elements are reachable (Line 4 in fig. 5), we mark it as the base point (Line 5 in fig. 5), and the network as the global sick network (Line 6 in fig. 5), add the network to the set h-PNs, and terminate the loop (Line 7 in fig. 5). Otherwise, the above operations are repeated, in the worst case, we need to repeat to the last vertex, and if the vertex is not the base point, we judge the network to be a non-global pathogenesis network. We perform the same decision operation on all candidate disease-causing networks, and then obtain the corresponding overall disease-causing network set h-PNs (Line 2-9 in FIG. 5).
Since the sick network is unweighted, and all edges of the network are given a weight of 1, it is clear that calculating reachability to all other points for a particular point is equivalent to calculating shortest paths to all other points for the particular point. The shortest path (Line 3 in fig. 5) was calculated in our study using a well-known and computationally efficient algorithm in graph theory (Dijkstra algorithm). However, although Dijkstra's algorithm is highly computationally efficient, the shortest path problem itself is a problem of particularly high computational complexity in graph theory. The computational complexity of Line 2-9 in fig. 5 reaches the order of one horror O (n × n)2) Where n represents the number of pathologists in a given network, O (n) corresponds to the complexity of Line 2 in FIG. 5, O (n)2) Representing the complexity of the Dijkstra algorithm. Worse still, the complexity only involves one network, and assuming we have 10 symptoms and 20 candidate disease-causing elements for each symptom, step S1 will generate 2010A candidate pathologist network. It is clear that this amount of computation is impossible to accomplish even for very well performing computers.
To solve the problem of computational efficiency of ALGORITHM 2To this end, we propose ALGORITHM 3 to optimize it, as shown in fig. 6. From a graph theory perspective, integrity (global property of the network) is highly correlated with vertex power (local property). We now propose the following three important theoretic reasoning about these two attributes, where NGRepresenting the number of pathologists with an income of 0 in the network.
1. Theorem 1: if G is an NGG is a non-integral network if the network is more than or equal to 2.
And (3) proving that: if not, G is global, then there must be at least one base point in G, not denoted v, we assert: v ∈ root (G) holds. Because of NGIs more than or equal to 2, and therefore, v is inevitable1，v2E root (G). If it is
Figure BDA0002931930360000119
Then
Figure BDA0002931930360000111
Is contradictory with v as the base point. On the other hand, because v ∈ root (G) and NGNot less than 2, therefore, v ≠ v' epsilon root (G) necessarily, and
Figure BDA0002931930360000112
if not this contradicts v' ∈ root (G). So v cannot be the base point for G.
In conclusion, propositions hold.
2. Theorem 2: if G is an NGNetwork of 1 and vrIs the vertex with an in-degree of 0, G is the global network
Figure BDA0002931930360000113
vrIs a base point.
And (3) proving that:
Figure BDA00029319303600001110
proof by theorem 1, N G1 and G is the integrity network, we can know that for vrE root (G), v must be the base point of G.
Figure BDA00029319303600001111
By definition, G is the integrity network.
In conclusion, propositions hold.
Definition 2: in the network G, for any two roads p1∈P(v，v′)，p2∈P(r，r′)If v' ═ r, i.e. p1Has an end point of p2Then we can connect these two roads to get P(v，v′)Let us denote the road so connected as p2*p1。
Introduction 1: definition of
Figure BDA0002931930360000114
Obviously, we have
Figure BDA00029319303600001112
Thus R is a binary relationship over G. Further, R is an equivalent relationship on G. If (x, y) is ∈ R, we will remember x to y.
And (3) proving that: self-reflexibility: for the
Figure BDA0002931930360000115
Always have
Figure BDA0002931930360000116
Thus x to x.
Symmetry: let x to y, i.e
Figure BDA0002931930360000117
Thus, it is possible to provide
Figure BDA0002931930360000118
Therefore, y to x.
Transferability: let x-y, y-z, then
Figure BDA0002931930360000121
Let us take pxy∈P(x，y)，pyz∈P(y，z)，pyx∈P(y，x)，pzy∈P(z，y). By definition of the link (see definition 2), we have pyz*pxy∈P(x，z)And p isyx*pzy∈P(z，x)Thus, therefore, it is
Figure BDA0002931930360000122
Further, x to z are present.
In summary, R is the equivalence relation on G. We call G/R the quotient network of G.
2, leading: (polymerization theory) G is the entire network, x, y ∈ VGAnd x to y. Then if x is the base point
Figure BDA0002931930360000123
y is the base point. That is, in the equivalence class divided by the equivalence relation R, the vertices in the same equivalence class are either all base points or are not all base points.
And (3) proving that:
Figure BDA0002931930360000124
since x is the base point, then
Figure BDA0002931930360000125
From the nature of the base point
Figure BDA0002931930360000126
Therefore we get
Figure BDA0002931930360000127
Due to x to y, therefore
Figure BDA0002931930360000128
Get
Figure BDA0002931930360000129
Then there is always
Figure BDA00029319303600001210
So y is the base point.
Figure BDA00029319303600001211
Propositions are symmetric and therefore true.
In conclusion, the reasoning holds.
3. Theorem 3: (network invariant theorem) if the network G is global
Figure BDA00029319303600001212
Its merchant network is monolithic.
And (3) proving that: from lemma 1, we know that for any network G, its quotient network G/R must exist.
Figure BDA00029319303600001217
Considering the natural projection: pi: g → G/R
Take a base point G and a road $ p $inG, note [ G $)]The equivalence class obtained for g, [ p ]]For the equivalence class obtained for a road p, using R as an equivalence relation and the definition of R, for any road pv∈P(g，v)Is a group ofv]∈P([g]，[v]). And then to
Figure BDA00029319303600001213
Is provided with
Figure BDA00029319303600001214
Thus [ g ]]Is the base point of G/R, so G/R is the whole network.
Figure BDA00029319303600001215
Consider pull-back: pi-1：G/R→G
Wherein, for
Figure BDA00029319303600001216
From the definition of pullback, one can know pi-1([g])：＝{v∈VGπ(v)＝[g]}. Taking a base point [ G ] in G/R]Remember Ug＝π-1([g]) It is obvious that
Figure BDA0002931930360000131
Having x-y, and thus pi-1The vertices are well defined. Therefore, pi can be directly defined-1([g]) G. For a road [ p ] in any G/R]∈P([v]，[v′])To a
Figure BDA0002931930360000132
And
Figure BDA0002931930360000133
since there is always px∈P(x，v)，px′∈P(v，x)，py∈P(y，v′)，py′∈P(v′，y). Thus only P needs to be found(v，v′)Can be defined by-1([p])＝p∈P(v，v′)So that pi-1It is also well defined for roads. Thus, pair
Figure BDA0002931930360000134
We have a pi-1([pv])＝px∈P(g，x)Then get pxv∈P(x，v)Thus has p(x，v)*px∈P(g，v). Traversing all equivalence classes, so that G is the base point of G, and knowing from the aggregation theorem that the base point G is independent of the choice, G is the integral network
Inference 1: if G is an NGNetwork of 0, in which there must be a directed ring GsFrom GsGet the vertex v of the renrG is the global network
Figure BDA0002931930360000135
vrIs a base point.
And (3) proving that:
Figure BDA0002931930360000136
from the network invariant theorem, G is the global network, and therefore G/R is also the global network, and thus there must be a base point in G/R.
[g]And pi-1([g]) Is a directed ring G in GsThen, the polymerization theory is used to show that any point in the ring is the base point of G.
Figure BDA0002931930360000137
Due to vrIs the base point, by definition, G is the integrity network.
In conclusion, it is concluded that this is true.
The description of ALGORITHM 3 is then as follows: for each candidate pathogenesis network Gi(see Line 1 in FIG. 6), we count the number of disease elements whose degree of entry is 0
Figure BDA0002931930360000138
(Line 2 in FIG. 6).
Figure BDA0002931930360000139
The value of (A) is divided into three cases, the first case
Figure BDA00029319303600001310
According to theorem 1, GiIs a non-global pathogenesis network (Line 36 in fig. 6); second case
Figure BDA00029319303600001311
Namely, a disease element v with an introductivity of 0 existsrAccording to theorem 2, GiIs equal to the pathogenesis v for the integral pathogenesis networkrAs a base point, v can be judged by Dijkstra algorithmrAccessibility to other pathogenesis A (v)r) If A (v)r) Is true ("1"), then 1, v is according to definitionrIs a base point, GiTo the global pathogenesis network, GiIs added to G+(Line 714); third case
Figure BDA00029319303600001312
There must be a directed ring G in the networksFrom GsGet the pathogenesis v of the conception vesselrAccording to the inference 1, G is the whole network equivalent to vrIs a base point, we are for vrThe same operations as in the second case are performed and it is determined whether it is the global disease-causing network (lines 15-25 in fig. 6). It is evident that the algorithm complexity is O (1) in the first case and O (n) in the second case2) I.e., the complexity of the Dijkstra algorithm, and the third case is also the complexity O (n) of the Dijkstra algorithm2) Line 2-16 has an overall algorithm complexity of O (n)2). In summary, given any disease network, the complexity of the ALGORITHM for determining whether it is the global disease network is reduced from O (n3) (ALGORITHM 2) to O (n)2)(ALGORITHM 3)。
S3 establishment of clinical pathogenesis network
All the global disease networks, i.e., h-PNs, corresponding to the set of symptoms are obtained from step S2, however, the number of global disease networks is not unique. It is easy to understand that, for the same group of symptoms, different genres have different thinking of syndrome differentiation, and even different individuals in the same genre have different syndrome differentiation outcomes, and the individual syndrome differentiation is an important inherent property connotation of TCM. The graph theory is not only an important data analysis tool, but also an important data visualization tool. The network syndrome differentiation screens out a limited overall disease pathogenesis network from infinite candidate disease pathogenesis networks by means of data analysis of graph theory (step S2), visually displays the limited overall disease pathogenesis network by means of visualization of the graph theory, and screens out a unique clinical disease pathogenesis network by fusing individual experience of doctors based on human-computer interaction (step S3). The traditional Chinese medicine dialectical theory mode not only restores the inherent overall concept of the traditional Chinese medicine dialectical process in a digitalized way, but also follows the inherent individualized dialectical characteristics of the traditional Chinese medicine dialectical process, and the traditional Chinese medicine dialectical mode is well-known and innovated.
Example 2
As shown in fig. 7, a network-based syndrome differentiation and pathogenesis-based traditional chinese medicine syndrome differentiation system (e.g., a computer server with program execution function) according to an exemplary embodiment of the present invention includes at least one processor, a power source, and a memory and an input/output interface communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method disclosed in any one of the preceding embodiments; the input and output interface can comprise a display, a keyboard, a mouse and a USB interface and is used for inputting and outputting data; the power supply is used for supplying electric energy to the electronic equipment.
Preferably, the system further comprises a human-computer interaction module, which can be a touch screen or other input and output devices, for implementing the human-computer cooperation strategy.
Those skilled in the art will understand that: all or part of the steps for realizing the method embodiments can be completed by hardware related to program instructions, the program can be stored in a computer readable storage medium, and the program executes the steps comprising the method embodiments when executed; and the aforementioned storage medium includes: various media that can store program codes, such as a removable Memory device, a Read Only Memory (ROM), a magnetic disk, or an optical disk.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.

Claims (7)

1. A traditional Chinese medicine pathogenesis network construction method based on network syndrome differentiation is characterized by comprising the following steps:
s1, generating a set PNs of all candidate pathogenesis networks according to all the symptom-pathogenesis binary relations and pathogenesis-pathogenesis binary relations;
s2, according to the accessibility of one pathogenesis element in the candidate pathogenesis network to all other pathogenesis elements, the overall property of the pathogenesis network is judged, the same judgment operation is carried out on all candidate pathogenesis networks, and the limited set h-PNs of the overall pathogenesis network is screened out from the set PNs.
2. The method for constructing a network-based syndrome differentiation traditional Chinese medicine pathogenesis network according to claim 1, wherein in step S1, all the symptom-pathogenesis binary relations and pathogenesis-pathogenesis binary relations are respectively stored in matrix C and matrix M, a set of clinical random symptom group S is given, and the algorithm for generating all candidate pathogenesis networks corresponding to S specifically comprises:
s11 reading symptoms S from matrix C respectivelyiCorresponding candidate pathogenesis elements are respectively stored in the set CiIn respect of all CiExecuting Cartesian product operation, and exhaustively enumerating to obtain all disease pathogenesis element sets V, i.e. generating
Figure FDA0002931930350000011
Set of candidate disease pathogenesis, symptom SiHas | CiSelecting | species;
s12 reading V from matrix MiAll corresponding disease mechanism chains are stored to a disease mechanism chain set Ei(ii) a Wherein the disease chain is a binary causal relationship between disease mechanism elements;
s13 integrating disease pathogenesis ViThe chain of pathogenesis EiCoupling to obtain all candidate pathogenesis networks, wherein the pathogenesis elements are collected ViSet of vertices, chains of pathogenesis E, representing the networkiRepresenting the edge of the network, and storing the edge into a set PNs to obtain the edge
Figure FDA0002931930350000012
And (5) planting candidate disease machine networks.
3. The method of claim 2, wherein the step S2 of determining any candidate disease mechanism network comprises:
starting to judge the accessibility of the disease element to all other disease elements from the first disease element of the network, if the disease elements are all accessible, marking the disease elements as base points, marking the network as an integral disease network, adding the network to the set h-PNs and terminating the cycle; if not, the above operations are repeatedly executed to judge the next pathogenesis element, and if the last vertex is not the base point, the network is judged to be a non-integral pathogenesis network.
4. The method for constructing a network-based syndrome differentiation pathogenesis network of traditional Chinese medicine according to claim 3, wherein said step S2 further comprises: for each candidate pathogenesis network GiAnd judging according to the number of pathogenesis elements with the degree of incidence of 0:
if it is
Figure FDA0002931930350000021
Then G isiA non-global pathogenesis network;
if it is
Figure FDA0002931930350000022
V is judged by Dijkstra algorithmrAccessibility to other pathogenesis A (v)r) If A (v)r) Is true, then vrIs a base point, GiTo the global pathogenesis network, GiIs added to G+；
If it is
Figure FDA0002931930350000023
There is a directed ring G in the networksFrom GsGet the pathogenesis v of the conception vesselrV is judged by Dijkstra algorithmrAccessibility to other pathogenesis A (v)r) If A (v)r) Is true, then vrIs a base point, GiTo the global pathogenesis network, GiIs added to G+。
5. The method for constructing a network-based syndrome differentiation pathophysiological network of traditional Chinese medicine according to any one of claims 1-4, further comprising:
s3 adopts man-machine cooperation strategy, network syndrome differentiation determines the only individual disease network from the limited overall disease network.
6. A traditional Chinese medicine syndrome differentiation system based on a network syndrome differentiation pathogenesis network is characterized by comprising at least one processor and a memory which is in communication connection with the at least one processor; the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 5.
7. The system of claim 6, further comprising a human-computer interaction module.
CN202110149996.2A 2021-02-03 2021-02-03 Traditional Chinese medicine pathogenesis network construction method and system based on network syndrome differentiation Active CN112786197B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110149996.2A CN112786197B (en) 2021-02-03 2021-02-03 Traditional Chinese medicine pathogenesis network construction method and system based on network syndrome differentiation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110149996.2A CN112786197B (en) 2021-02-03 2021-02-03 Traditional Chinese medicine pathogenesis network construction method and system based on network syndrome differentiation

Publications (2)

Publication Number Publication Date
CN112786197A true CN112786197A (en) 2021-05-11
CN112786197B CN112786197B (en) 2023-08-18

Family

ID=75760770

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110149996.2A Active CN112786197B (en) 2021-02-03 2021-02-03 Traditional Chinese medicine pathogenesis network construction method and system based on network syndrome differentiation

Country Status (1)

Country Link
CN (1) CN112786197B (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103455723A (en) * 2013-09-02 2013-12-18 成都中医药大学 Building method of traditional Chinese medicine intelligent syndrome differentiation medical knowledge model based on ISO-R rule
CN106503789A (en) * 2016-11-08 2017-03-15 西安电子科技大学宁波信息技术研究院 Loop-free shortest path searching method based on Di Jiesitela and minimax ant colony
CN106650253A (en) * 2016-12-16 2017-05-10 新绎健康科技有限公司 Dialectical system of traditional Chinese medicine
CN106933994A (en) * 2017-02-27 2017-07-07 广东省中医院 A kind of core disease card relation construction method based on knowledge of TCM collection of illustrative plates
CN107292091A (en) * 2017-06-05 2017-10-24 成都元峰科技有限公司 A kind of construction method of traditional Chinese medical science syndrome elements differentiation medical knowledge model
CN108986911A (en) * 2018-07-06 2018-12-11 成都中医药大学 A kind of differential diagnosis in tcm opinion controls data processing method
CN111462824A (en) * 2020-06-04 2020-07-28 东北大学 Reachable probability query method for gene regulation and control network

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103455723A (en) * 2013-09-02 2013-12-18 成都中医药大学 Building method of traditional Chinese medicine intelligent syndrome differentiation medical knowledge model based on ISO-R rule
CN106503789A (en) * 2016-11-08 2017-03-15 西安电子科技大学宁波信息技术研究院 Loop-free shortest path searching method based on Di Jiesitela and minimax ant colony
CN106650253A (en) * 2016-12-16 2017-05-10 新绎健康科技有限公司 Dialectical system of traditional Chinese medicine
CN106933994A (en) * 2017-02-27 2017-07-07 广东省中医院 A kind of core disease card relation construction method based on knowledge of TCM collection of illustrative plates
CN107292091A (en) * 2017-06-05 2017-10-24 成都元峰科技有限公司 A kind of construction method of traditional Chinese medical science syndrome elements differentiation medical knowledge model
CN108986911A (en) * 2018-07-06 2018-12-11 成都中医药大学 A kind of differential diagnosis in tcm opinion controls data processing method
CN111462824A (en) * 2020-06-04 2020-07-28 东北大学 Reachable probability query method for gene regulation and control network

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
许强: ""基于有向图构建张之文老师关于咳嗽病的计算机智能辨证模型的方法学研究"" *
许强等: ""网络辨证的算法原理"" *
许强等: ""通过有向图构建证素的组合法则—ISO -Ｒ 法则"" *

Also Published As

Publication number Publication date
CN112786197B (en) 2023-08-18

Similar Documents

Publication Publication Date Title
Holzinger et al. Knowledge discovery and interactive data mining in bioinformatics-state-of-the-art, future challenges and research directions
Holzinger et al. Knowledge discovery and data mining in biomedical informatics: The future is in integrative, interactive machine learning solutions
Cohen Computer algebra and symbolic computation: Mathematical methods
Parsons et al. Adjustable properties of visual representations: Improving the quality of human‐information interaction
Steen et al. Extensional higher-order paramodulation in Leo-III
US10585902B2 (en) Cognitive computer assisted attribute acquisition through iterative disclosure
WO2018149930A1 (en) Neurological data processing
Ravuri et al. Learning from the experts: From expert systems to machine-learned diagnosis models
Sheremet et al. A logic for concepts and similarity
Barendregt et al. Electronic communication of mathematics and the interaction of computer algebra systems and proof assistants
Subbaroyan et al. Minimum complexity drives regulatory logic in Boolean models of living systems
Geniesse et al. NeuMapper: A scalable computational framework for multiscale exploration of the brain’s dynamical organization
CN112786197A (en) Traditional Chinese medicine pathogenesis network construction method and system based on network syndrome differentiation
Mahboubi Implementing the cylindrical algebraic decomposition within the Coq system
Fu et al. Formalization of the Equivalence among Completeness Theorems of Real Number in Coq
Christou et al. Identifying all abelian periods of a string in quadratic time and relevant problems
Krüger Bitvectors in SMT-RAT and their application to integer arithmetics
Batzoglou Gödel's Incompleteness Theorem
Kodytek et al. Automated code development based on genetic programming in graphical programming language: A pilot study
Sultana et al. Querying KEGG pathways in logic
Barzdins et al. Fast Ad Hoc Queries Based on Data Ontologies.
Monnin et al. Matching pharmacogenomic knowledge: particularities, results, and perspectives
Berenbeim Foundations for the Analysis of Surreal-valued genetic functions
CN110827989B (en) Control method for processing medical data based on key factors
Martínez Applications of Foundational Proof Certificates in theorem proving

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant