Deep venous thrombosis risk assessment ruler and assessment system

Abstract

The invention discloses a deep vein thrombosis risk assessment ruler and assessment system, and relates to the technical field of thrombosis assessment, and the method comprises the following steps: S1, constructing a DVT risk factor structured mapping unit; s2, constructing a hierarchical risk feature tensor by using the risk factor set; s3, constructing an interactive risk field pedigree matrix by using the risk feature tensor; s4, multi-channel standard structure recombination is carried out by using the pedigree matrix set; s5, constructing an interactive risk field pedigree matrix by using the general evaluation label set; and S6, performing individual risk path reconstruction and action mechanism visual expression based on the risk field pedigree matrix. By setting a multi-scale interactive structure factor identification mechanism and an input structure label set linkage compression process, the method not only introduces a structure channel identification and local label projection mechanism in an input stage, but also dynamically constructs a pedigree matrix; and the propagation path coupling between the input factors and the adaptation relationship between the tag systems are obviously enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of thrombosis assessment, and in particular to a deep vein thrombosis risk assessment scale and an assessment system. Background Art

[0002] Deep vein thrombosis (DVT) is a common clinical vascular disease with a high risk of thrombosis. Its risk assessment is directly related to the effectiveness of clinical early warning, intervention, and tiered management. Currently, various DVT risk assessment methods based on scoring rules, static indicator classification, and limited pathway modeling exist. However, these methods generally suffer from insufficient response to dynamic interaction factors, partial understanding of risk transmission pathways, and poor adaptability of labeling results. This makes it difficult to achieve accurate modeling and dynamic prediction of individual differences in complex scenarios.

[0003] In existing technologies, the processing of input features mostly remains at the stage of static single-scale feature screening and subjective experience weight scoring. It lacks a substructure analysis mechanism for "structural-level risk propagation" and is difficult to support the deep evolutionary reasoning of subsequent pedigree modeling. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the present invention provides a deep vein thrombosis risk assessment scale and an assessment system to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: In a first aspect, an embodiment of the present invention provides a deep vein thrombosis risk assessment scale and assessment system, comprising the following steps: S1. Construct a DVT risk factor structured mapping unit to obtain a risk factor set; S2. constructing a hierarchical risk feature tensor using the risk factor set to obtain a risk feature tensor; S3. Use the risk feature tensor to construct the interactive risk field spectrum matrix and obtain the spectrum matrix set; S4. Use the pedigree matrix set to perform multi-channel standard structure reorganization to obtain a universal evaluation label set; S5. Use the universal evaluation label set to construct an interactive risk field spectrum matrix to obtain a risk field spectrum matrix; S6. Based on the risk field spectrum matrix, individual risk paths are reconstructed and the action mechanism is visualized to achieve risk deduction output under multiple labels.

[0006] To further optimize this technical solution, step S1 obtains factors from the patient information system, and performs unified expression and splicing to obtain the following factor expression: ; in, Indicates the The standardized value of each risk factor; It is a structured input set with a unified format and meaning, and its length is 13 dimensions.

[0007] Further optimizing this technical solution, the step S2 firstly converts the factor action mechanism, medical pathway and risk dimension into Divided into several subdomains , such that: ; Each subset Indicates that it belongs to a subdomain A subset of factors.

[0008] To further optimize this technical solution, step S2 recombines the factors in each subdomain into a vector ; ; in, Indicates the The factor vector in the factor subdomain has a length of .

[0009] Further optimizing this technical solution, the step S2 integrates the subdomain vectors into a three-dimensional tensor structure to form a risk feature tensor after the stratification is completed. ; ; in, : No. Factor subvectors under subdomains; : The number of subdomains divided; : the length of each subdomain vector, ; : Represents a structural organization mapping function that assembles factors into tensors in subdomain order.

[0010] Further optimizing this technical solution, the step S3 is to In the above example, we construct the internal risk spectrum matrix: ; in, Representing semantic subdomains Middle The first and The strength of the interaction between risk factors; Representing semantic subdomains Middle The eigenvalue of each risk factor; is the measurement function of the interaction relationship between risk factors.

[0011] To further optimize this technical solution, the risk factor interaction relationship measurement function in step S3 is as follows: ; in, The risk factor eigenvector representing the two risk factors whose correlations are to be calculated; Represents a vector and The covariance between is a vector The standard deviation of is a vector The standard deviation of Factor and Relevance of risk roles within the current subdomain.

[0012] To further optimize this technical solution, the risk field spectrum generation function of the response in step S5 is: ; in, : No. The structural expression vector of each sample; : No. The labels of the samples; : Label level conversion function, used to assign risk field strength between structures under different labels; : Normalization constant used to keep the range of the pedigree matrix controllable.

[0013] To further optimize this technical solution, step S5 includes the following sub-processes during execution: Constructing label weight mapping function : Construct an enhanced diagonal weight matrix for the “high risk” label; Construct a neighbor weak connection matrix for the "medium risk" label; Construct a sparse diagonal matrix for the “low risk” label; Output mapping function set ; Perform interactive structure generation: , which captures the label-guided interaction patterns among structural dimensions in this sample; Full sample pedigree superposition and fusion: Use the average fusion mechanism to avoid the number of samples with a certain label dominating the pedigree results; Final Output , which represents the global risk structure action spectrum matrix.

[0014] To further optimize this technical solution, the present invention also includes a deep vein thrombosis risk assessment scale, which sets a number of preset identification areas on the surface of the patient's limbs and combines a dynamic data interface linked to the business module to achieve standardized quantification of individual physiological characteristics and form a structured original input tensor.

[0015] In a second aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of a deep vein thrombosis risk assessment scale and assessment system as described in the first aspect of the present invention are implemented.

[0016] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of a deep vein thrombosis risk assessment scale and assessment system as described in the first aspect of the present invention are implemented.

[0017] Compared with the existing technology, the present invention provides a deep vein thrombosis risk assessment scale and assessment system, which has the following beneficial effects: This deep vein thrombosis risk assessment scale and assessment system, by setting up a multi-scale interactive structural factor identification mechanism and an input structure label set linkage compression process, not only introduces a structural channel identification and local label projection mechanism in the input stage, but also significantly enhances the coupling of propagation paths between input factors and the adaptation relationship between label systems through the dynamic construction of the pedigree matrix, thereby effectively solving the core problem of insufficient reasoning ability caused by the flat input factor structure and the fragmented label system in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 This is a flow chart of a deep vein thrombosis risk assessment system proposed by the present invention; Figure 2 A schematic diagram of the process of constructing an interactive risk field spectrum matrix for a deep vein thrombosis risk assessment system proposed in the present invention; Figure 3 This is a schematic diagram of the process of constructing an interactive risk field spectrum matrix for a deep vein thrombosis risk assessment system proposed in the present invention. DETAILED DESCRIPTION

[0020] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0021] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0022] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it constitute a separate or selective embodiment that is mutually exclusive with other embodiments.

[0023] Example 1: Reference Figures 1 to 3 , which is the first embodiment of the present invention, provides a deep vein thrombosis risk assessment system, comprising the following steps: S1. Construct a DVT risk factor structured mapping unit to obtain a risk factor set; Step S1 first obtains factors from the patient information system based on existing models (including the Caprini Score Model and the Padua Risk Score Model) and medical consensus, including: age, body mass index (BMI), previous history of VTE, malignancy, heart failure status, hereditary coagulation disorders, surgery duration, central venous catheter placement, steroid use, mobility, bed rest time, infection status, and history of lower limb trauma or recent fracture; Age, body mass index, duration of surgery, and bed rest time were then standardized using the Z-score method. Activity, infection status, malignancy, and heart failure status were processed using a hierarchical mapping method. Previous history of ventricular embolism (VTE), central venous catheterization, steroid use, hereditary coagulation disorders, and history of lower limb trauma or recent fracture were directly mapped to 0 or 1 using a mapping method, thereby achieving a unified expression of factors from different sources and units. Finally, the above normalized factors are concatenated into a fixed-length structured vector to obtain the following factor expression: ; in, Indicates the The standardized value of each risk factor; It is a structured input set with a unified format and meaning, and its length is 13 dimensions.

[0024] S2. constructing a hierarchical risk feature tensor using the risk factor set to obtain a risk feature tensor; Step S2: Unify the standardized risk factor set On this basis, a tensor structure with medical semantics and factor scope hierarchical logic is constructed ; First, according to the mechanism of action of factors, medical pathways and risk dimensions, Divided into several subdomains , such that: ; Each subset Indicates that it belongs to a subdomain A subset of factors.

[0025] Then the factors in each subdomain are recombined into vectors ; ; in, Indicates the The factor vector in the factor subdomain has a length of .

[0026] Finally, the subdomain vectors are integrated into a three-dimensional tensor structure to form the risk feature tensor after stratification. ; ; in, : No. Factor subvectors under subdomains; : The number of subdomains divided; : the length of each subdomain vector, ; : Represents a structural organization mapping function that assembles factors into tensors in subdomain order.

[0027] Most existing DVT risk assessment methods use linear scoring or factor-based weighting methods, such as the Caprini assessment method, which directly maps each factor into a numerical weight and sums them up without hierarchical modeling of the semantic structure or interaction relationship between factors. This step uses the structure mapping function Constructing a tensor structure with structural hierarchy and semantic domain distinction It not only has structural alignment, but also provides an explicit multi-domain hierarchical input basis for subsequent risk pathway modeling, which is different from the existing one-way addition or linear regression input mechanism.

[0028] S3. Use the risk feature tensor to construct the interactive risk field spectrum matrix and obtain the spectrum matrix set; Step S3 uses the risk feature tensor of step S2 , construct a pedigree matrix that expresses the coupling relationship between risk factors within the semantic subdomain , used to reveal the mutual influence and action paths among the 13 risk factors in each semantic subdomain; In each semantic subdomain In the above example, we construct the internal risk spectrum matrix: ; in, Representing semantic subdomains Middle The first and The strength of the interaction between risk factors; Representing semantic subdomains Middle The eigenvalue of each risk factor; is the measurement function of the interaction relationship between risk factors, and its function formula is as follows: ; in, The risk factor feature vectors representing the two risk factor correlations to be calculated (from the same subdomain); Represents a vector and The covariance between is a vector The standard deviation of is a vector The standard deviation of Factor and relevance of risk roles within the current subdomain; The tensor form of the step output is: .

[0029] Step S3, during the construction of the interactive risk field spectrum matrix, i.e., when the risk spectrum matrix is ​​used in practice, includes the following processes: First, extract the factor vectors within the same semantic subdomain from the tensor: Input the risk feature tensor output from step S2 , for each semantic subdomain , extract the 13 factor eigenvectors under this subdomain , corresponding to the input items in the risk spectrum matrix formula; Then for each semantic subdomain , and calculate the correlation function between the 13 factors within it: ; Finally, we get the pedigree matrix set: ; Corresponding to the risk spectrum matrix formula ; Finally, the pedigree matrix is ​​transformed into a risk effect diagram. Mapping to semantic subdomains The risk factor network graph below is calculated to complete all Next , and obtain the complete pedigree graph structure for subsequent graph analysis.

[0030] Most existing methods are based on static factor weight evaluation, such as logistic regression, weighted scoring or principal component analysis (PCA). The interactive risk field spectrum matrix constructed in this step uses semantic subdomains as units to quantify the risk propagation relationship between factors within the same subdomain, which belongs to structured graphical modeling. The correlation function is context-dependent rather than a fixed global weight, and therefore has the ability to dynamically map risk states.

[0031] Different from traditional single-factor or weighted evaluation methods, this step can provide a structural basis for subsequent path prediction, heterogeneous risk propagation modeling, etc. It has the ability of graph reasoning and network structure recognition, and is more suitable for complex medical risk modeling scenarios with multi-factor interactions.

[0032] S4. Use the pedigree matrix set to perform multi-channel standard structure reorganization to obtain a universal evaluation label set; Step S4 first performs the standard structural alignment of the pedigree matrix set, and calculates the graph Laplacian matrix of each pedigree matrix through the mature Laplacian Eigenmaps method, obtains its eigenvector as a low-dimensional representation, and obtains each The corresponding embedding vector set , keep the unified embedding dimension consistent, and obtain a multi-channel unified embedding representation set ; Then, multi-channel structure reorganization and fusion are performed, and the set of channel embedding vectors is structurally fused to construct a unified risk spectrum expression. The rank reduction of the multi-channel embedding tensor is performed through the mature CANDECOMP / PARAFAC (CP) tensor decomposition method to obtain the standard structure expression matrix ,in represents the number of evaluation samples, is the structural dimension after fusion; Finally, a general evaluation label set is extracted. For unlabeled data sets, K-means is used to divide risk levels based on structural expressions. For labeled historical data, support vector machines (SVM) are used for label mapping learning to ultimately form a general evaluation label set: .

[0033] S5. Use the universal evaluation label set to construct an interactive risk field spectrum matrix to obtain a risk field spectrum matrix; Step S5 is based on the general evaluation label set output by step S4 , a risk field pedigree matrix model is constructed to simulate and predict the interaction path of multiple risk factors. The corresponding risk field pedigree generating function is: ; in, : No. The structural expression vector of each sample; : No. The labels of the samples; : Label level conversion function, used to assign risk field strength between structures under different labels; : Normalization constant used to keep the range of the pedigree matrix controllable.

[0034] In step S5, a risk field spectrum generation function is constructed to derive potential interaction path patterns between structural factors at different risk levels. The use of the function includes: According to the label Introduce the corresponding transformation matrix ; For each sample structure vector Multiply left and right , extracting label-guided structural interaction mapping; All samples are weighted and superimposed by the normalization factor Get the final pedigree matrix , whose dimensions are , which represents the spectrum of interactions between structural dimensions.

[0035] During the execution of step S5, the following sub-processes are included: Constructing label weight mapping function : Construct an enhanced diagonal weight matrix for the “high risk” label; Construct a neighbor weak connection matrix for the "medium risk" label; Construct a sparse diagonal matrix for the “low risk” label; Output mapping function set ; Perform interactive structure generation: , which captures the label-guided interaction patterns among structural dimensions in this sample; Full sample pedigree superposition and fusion: Use the average fusion mechanism to avoid the number of samples with a certain label dominating the pedigree results; Final Output , which represents the global risk structure action spectrum matrix.

[0036] Compared with existing mature technologies, step S5 has achieved significant breakthroughs in label modeling, structural interaction modeling, and output expression: by introducing the label mapping function Implementing Tags Pair structure vector The pedigree effect breaks the limitation of independent or weighted processing of each label in traditional methods; further by constructing a bilateral mapping expression , realize the bidirectional interaction modeling of structural dimensions under label control, and form a high-order cross-spectrum; the final output is the interaction spectrum matrix between structures , not only has structural a priori significance, but also supports subsequent risk path deduction and visualization, which is different from the discrete labels or risk scores output by traditional methods.

[0037] S6. Reconstruct individual risk paths and visualize the mechanism of action based on the risk field spectrum matrix to achieve multi-label risk deduction output; Step S6 includes the following sub-processes in the process of reconstructing individual risk pathways and visualizing the mechanism of action: Slice localization and channel separation of individual risk field lineages: Through label condition retrieval mechanism (such as using conditional logic filters, or through Tensor slicing operations), The action spectrum of the target individual under the target label set is located in the target label set; If the target individual is a time-related entity (e.g., hospitalized patients), the time series indexing method is used for segmentation; Path causal chain construction and node attribution: Adopting a directed graph modeling mechanism, risk spectrum data is converted into node representation (indicating factors or structural positions) and edge representation (indicating risk impact direction or deductive relationship); Causal graph modeling tools (such as Bayesian Network or Granger causal inference) can be used to identify potential path causal chains; For static labels, graph traversal algorithms (such as DFS / BFS) can be used to complete the shortest chain construction of causal paths; For dynamic labels, event flow analysis tools (such as Petri nets) can be introduced to model dynamic evolution paths; Risk mechanism map generation and label deduction path summary: Based on the path diagram structure, a three-dimensional mechanism map of "structural factor-label-path influence" is formed; Cluster and merge multiple paths with the same label to generate a representative "typical deduction path subset"; Path similarity measurement tools (such as Graph Edit Distance) can be used to reduce path evolution; Visual generation and output of risk deduction results: Convert the generated path mechanism graph into a visual graph and use graph structure rendering tools (such as D3.js and Graphviz) to dynamically draw nodes and edges; Add label highlighting and importance assessment layers to path nodes to enhance visual interpretation; The deduction results can be output in the form of "label risk intensity-time series diagram", "path trajectory diagram", "individual comparison diagram", etc.

[0038] The present invention also includes a deep vein thrombosis risk assessment scale. This scale serves as the system's front-end data acquisition and structure recognition tool, featuring multi-channel, physiological structure mapping, and label projection capabilities. By defining several preset recognition areas on the patient's limb surface and integrating a dynamic data interface linked to the service module, the scale achieves standardized quantification of individual physiological characteristics, forming a structured raw input tensor. This provides fundamental data support for risk factor extraction in step S1 and interactive risk structure reconstruction in step S2.

[0039] The scale's functionality extends beyond physiological measurements. It also incorporates a risk label mapping and encoding module, dynamically generating the initial label state for pathway deduction based on historical case statistics and a label set compression model (step S4). Furthermore, its output tensor can be directly adapted to the pedigree matrix construction process (step S3), providing key input for risk propagation pathway modeling (step S5) and personalized risk mechanism visualization (step S6). The scale's overall structure is highly compatible with the system's subsequent analysis processes, making it a key entry point for transitioning from raw features to mechanism-level deduction.

[0040] Example 2: Reference Figures 1 to 3 , which is the second embodiment of the present invention, and provides a dynamic deduction of individual risk based on a pedigree matrix; This embodiment has the following implementation process: Dynamically collect risk factors and form time series input; Continuously collect a single object at different times The risk factor vector on (The dimensions are consistent with the main process), forming an observation time series matrix , which is the temporal risk trajectory sample set of the current object.

[0041] Introducing the pedigree matrix as a risk communication structure; Directly call the pedigree matrix output by step S3 of the main process , which is considered as a structural prior for the spatial propagation graph of risk evolution. This matrix contains the risk interaction relationships between different subdomains (such as wards, blood flow segments, clinical scenarios, etc.), which is used to constrain and guide the propagation path and transmission intensity of risk factors.

[0042] Construct a risk evolution model based on structural propagation; Will Input at each moment , in the structural constraint matrix Mapping and dynamic path reconstruction are performed to form a propagation-guided risk state in a continuous time series.

[0043] Real-time output of risk prediction and path visualization; Based on the above propagation results, the system can Give the current risk level , and combined with the propagation path to generate real-time visualization results, including risk diffusion direction, core risk subdomains, highly sensitive channels and other information, for clinical decision-making assistance or early intervention.

[0044] This real-time embodiment does not need to rebuild the main process model, by calling the pedigree structure established in step S3 , logically realized the following innovations: Achieve risk tracking with higher granularity from the perspective of individual time series; Keep the structure unchanged and only replace the timing input to reduce the system calculation burden; Dynamic propagation is based on a pedigree matrix rather than a reconstruction model, improving timeliness and interpretability; It is suitable for lightweight deployment (such as bedside monitoring systems, mobile terminals, etc.), facilitating practical application and promotion.

[0045] Example 3:

[0046] This embodiment also provides a computer device suitable for a deep vein thrombosis risk assessment scale and assessment system, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement a deep vein thrombosis risk assessment scale and assessment system as proposed in the above embodiments.

[0047] This embodiment further provides a storage medium having a computer program stored thereon. When the program is executed by a processor, a deep vein thrombosis risk assessment scale and assessment system as proposed in the above embodiments are implemented.

[0048] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0049] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0050] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0051] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting, or processing it in another suitable manner as necessary, and then storing it in a computer memory.

[0052] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having logic gate circuits for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc.

[0053] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A deep vein thrombosis risk assessment system, characterized in that: The following steps are involved: S1. Construct a DVT risk factor structured mapping unit to obtain a risk factor set; S2. constructing a hierarchical risk feature tensor using the risk factor set to obtain a risk feature tensor; S3. Use the risk feature tensor to construct the interactive risk field spectrum matrix and obtain the spectrum matrix set; S4. Use the pedigree matrix set to perform multi-channel standard structure reorganization to obtain a universal evaluation label set; S5. Use the universal evaluation label set to construct an interactive risk field spectrum matrix to obtain a risk field spectrum matrix; S6. Based on the risk field spectrum matrix, individual risk paths are reconstructed and the action mechanism is visualized to achieve risk deduction output under multiple labels.

2. A deep vein thrombosis risk assessment system according to claim 1, characterized in that: The step S1 obtains factors from the patient information system, and performs unified expression and splicing to obtain the factor expression in the following form: ； in, Indicates the The standardized value of each risk factor; It is a structured input set with a unified format and meaning, and its length is 13 dimensions.

3. A deep vein thrombosis risk assessment system according to claim 1, characterized in that: The step S2 firstly divides the factor into Divided into several subdomains , such that: ； Each subset Indicates that it belongs to a subdomain A subset of factors.

4. A deep vein thrombosis risk assessment system according to claim 3, characterized in that: The step S2 recombines the factors in each subdomain into a vector ; ； in, Indicates the The factor vector in the factor subdomain has a length of .

5. A deep vein thrombosis risk assessment system according to claim 3, characterized in that: The step S2 integrates the subdomain vectors into a three-dimensional tensor structure to form a risk feature tensor after stratification. ; ； in, : No. Factor subvectors under subdomains; : The number of subdomains divided; : the length of each subdomain vector, ; : Represents a structural organization mapping function that assembles factors into tensors in subdomain order.

6. A deep vein thrombosis risk assessment system according to claim 1, characterized in that: In step S3, each semantic subdomain In the above example, we construct the internal risk spectrum matrix: ； in, Representing semantic subdomains Middle The first and The strength of the interaction between risk factors; Representing semantic subdomains Middle The eigenvalue of each risk factor; is the measurement function of the interaction relationship between risk factors.

7. A deep vein thrombosis risk assessment system according to claim 6, characterized in that: The risk factor interaction relationship measurement function in step S3 has the following function formula: ； in, The risk factor eigenvector representing the two risk factors whose correlations are to be calculated; Represents a vector and The covariance between is a vector The standard deviation of is a vector The standard deviation of Factor and Relevance of risk roles within the current subdomain.

8. A deep vein thrombosis risk assessment system according to claim 1, characterized in that: The risk field spectrum generation function in step S5 is: ； in, : No. The structural expression vector of each sample; : No. The labels of the samples; : Label level conversion function, used to assign risk field strength between structures under different labels; : Normalization constant used to keep the range of the pedigree matrix controllable.

9. A deep vein thrombosis risk assessment system according to claim 8, characterized in that: During the execution of step S5, the following sub-processes are included: Constructing label weight mapping function : Construct an enhanced diagonal weight matrix for the "high risk" label; Construct a neighbor weak connection matrix for the "medium risk" label; Construct a sparse diagonal matrix for the "low risk" label; Output mapping function set ; Perform interactive structure generation: , which captures the label-guided interaction patterns among structural dimensions in this sample; Full sample pedigree superposition and fusion: Use the average fusion mechanism to avoid the number of samples with a certain label dominating the pedigree results; Final Output , which represents the global risk structure action spectrum matrix.

10. A deep vein thrombosis risk assessment scale, constructed based on the deep vein thrombosis risk assessment system according to any one of claims 1 to 9, characterized in that: The assessment scale sets several preset identification areas on the surface of the patient's limbs and combines them with a dynamic data interface linked to the business module to achieve standardized quantification of individual physiological characteristics and form a structured original input tensor.

Images