Intelligent management system and method for medical health record information

Abstract

The present application proposes a medical health record information intelligent management system and method, comprising: constructing a medical three-dimensional network, mapping medical terms into three-dimensional coordinate nodes, and calculating node distance through co-occurrence matrix; according to the attributes of each patient, projecting the medical three-dimensional network to a two-dimensional plane to form an interactive two-dimensional cognitive map; based on the two-dimensional cognitive map, optimizing the record classification rules to generate the optimal record classification rule set; based on the optimal record classification rule set and the two-dimensional cognitive map, planning the optimal search path; performing record search operation on the two-dimensional cognitive map according to the optimal search path, and quickly positioning and obtaining the target record according to the record node sequence on the path. The present application constructs a three-dimensional network to integrate information and reflect term association, projects into a two-dimensional cognitive map, optimizes the classification rules, and plans the optimal path, and accurately finds the target record according to the optimal path, avoiding irrelevant records, greatly improving the search efficiency and accuracy.

Description

Technical Field

[0001] This invention relates to the field of medical and health record information management technology, and in particular to an intelligent management system and method for medical and health record information. Background Technology

[0002] With the continuous advancement of medical technology and the rapid development of medical informatization, the quantity and complexity of medical and health records have increased dramatically. These records contain a wealth of information on medical terms such as diseases, symptoms, and medications, serving as a crucial foundation for medical research, clinical diagnosis, and patient treatment. However, how to efficiently manage and utilize this medical and health record information remains an urgent problem to be solved.

[0003] Currently, medical and health record management mainly relies on manual paper records or simple electronic record management systems. These management methods have the following shortcomings in terms of record storage, retrieval, and utilization: 1. They mainly rely on keyword matching or simple category directory retrieval, making it difficult to consider the semantic relationships and actual connections between medical terms, resulting in incomplete and inaccurate search results; 2. The search interface is usually relatively simple, lacking intuitiveness and interactivity, making it difficult for users to quickly find the records they need; 3. They mainly focus on record storage and simple retrieval, lacking the ability to deeply mine medical knowledge and optimize rules; 4. When new medical terms or information emerge, traditional systems may require large-scale modifications and upgrades to support the management and retrieval of this new information, which is costly and time-consuming. Summary of the Invention

[0004] Therefore, the purpose of this invention is to propose an intelligent management system and method for medical and health record information to solve the problems mentioned above.

[0005] The intelligent management system for medical and health record information proposed according to the present invention includes:

[0006] 3D Network Module: Used to construct a medical 3D network to map medical terms to 3D coordinate nodes and calculate node distances through a co-occurrence matrix;

[0007] Two-dimensional map module: used to project the medical three-dimensional network onto a two-dimensional plane based on each patient's attributes, forming an interactive two-dimensional cognitive map;

[0008] Classification rules module: used to optimize the classification rules of archives based on two-dimensional cognitive maps and generate the optimal set of archive classification rules;

[0009] Search path module: Used to plan the optimal search path based on the optimal set of document classification rules and two-dimensional cognitive map;

[0010] Document retrieval module: Used to perform document retrieval operations on a two-dimensional cognitive map using the optimal retrieval path, and quickly locate and retrieve target documents based on the order of document nodes on the path.

[0011] This invention also proposes an intelligent management method for medical and health record information, used to implement the aforementioned intelligent management system for medical and health record information, the method comprising:

[0012] A medical 3D network is constructed to map medical terms to 3D coordinate nodes, and the node distance is calculated through a co-occurrence matrix;

[0013] Based on each patient's attributes, the medical 3D network is projected onto a 2D plane to form an interactive 2D cognitive map;

[0014] The document classification rules are optimized based on a two-dimensional cognitive map to generate the optimal document classification rule set;

[0015] Based on the optimal set of archive classification rules and a two-dimensional cognitive map, the optimal retrieval path is planned;

[0016] The optimal search path is used to perform file retrieval on a two-dimensional cognitive map, and the target file is quickly located and retrieved based on the order of file nodes on the path.

[0017] Furthermore, the step of optimizing the document classification rules based on the two-dimensional cognitive map to generate the optimal document classification rule set includes:

[0018] Encode all file classification rules into binary strings;

[0019] With retrieval efficiency and misdiagnosis avoidance rate as dual objectives, the application effect of archive classification rules on a two-dimensional cognitive map is simulated. The optimal set of archive classification rules is selected through multiple rounds of iteration to guide archive classification and retrieval.

[0020] New case data is injected automatically on a regular basis, triggering the mutation and crossover of the file classification rule set, eliminating inefficient rules, adjusting the file classification rule set, and obtaining the updated optimal file classification rule set.

[0021] Furthermore, the steps of optimizing retrieval efficiency and misdiagnosis avoidance rate with dual objectives, simulating the application effect of document classification rules on a two-dimensional cognitive map, and selecting the optimal set of document classification rules through multiple rounds of iteration include:

[0022] Generate initial rule individuals to form an initial population, where the binary string of each file classification rule represents a rule individual;

[0023] On a two-dimensional cognitive map, simulated archival data is classified according to archival classification rules;

[0024] Record the average path length required to retrieve a specific file, both when and after classifying it using file classification rules.

[0025] The path shortening rate is calculated using the formula: Path shortening rate = (Average path length without rules - Average path length with rules) / Average path length without rules × 100%;

[0026] The number of conflicting rules is counted, and the misdiagnosis avoidance rate is calculated using the formula: Misdiagnosis Avoidance Rate = (1 - Number of Conflicting Rules / Total Number of Rules) × 100%;

[0027] The path shortening rate and misdiagnosis avoidance rate are weighted and summed to obtain the overall fitness of each rule individual;

[0028] Select rule individuals with high overall fitness to enter the next generation, and perform crossover and mutation operations to obtain new rule individuals. Repeat the iteration until the overall fitness converges, then terminate the iteration and select the rule individual with the highest overall fitness as the optimal file classification rule.

[0029] Furthermore, the step of planning the optimal retrieval path based on the optimal set of document classification rules and the two-dimensional cognitive map includes:

[0030] Determine the file association strength value;

[0031] The strength of file associations is converted into pheromone concentration, and a pheromone matrix is ​​constructed.

[0032] Based on the pheromone matrix and the optimal set of file classification rules, the ant colony releases pheromones on a two-dimensional cognitive map to perform retrieval, and plans the optimal retrieval path through a probabilistic selection formula.

[0033] Furthermore, the steps of simulating ant colonies releasing pheromones on a two-dimensional cognitive map for retrieval based on the pheromone matrix and the optimal file classification rule set, and planning the optimal retrieval path using a probabilistic selection formula, include:

[0034] For each ant in the colony, starting from the initial file node of the two-dimensional cognitive map, the next file node to be visited is selected according to the probability selection formula, and the retrieval path from the initial file to the target file is recorded.

[0035] After all ants have completed one path construction, the pheromone matrix is ​​evaporated.

[0036] For each path constructed by an ant, pheromones are released based on the quality of the path.

[0037] Update the pheromone matrix and add the pheromones released by all ants to the corresponding edges;

[0038] Repeat the iteration until the pheromone matrix tends to stabilize. After the iteration ends, select the path with the highest pheromone concentration as the optimal retrieval path based on the current pheromone matrix and the probability selection formula.

[0039] Furthermore, after the step of planning the optimal retrieval path using the probabilistic selection formula, the method further includes:

[0040] Based on real-time medical data and the current pheromone matrix, the pheromone concentration is adjusted according to medical needs to obtain an adjusted pheromone matrix, making the retrieval path more in line with current medical needs.

[0041] Furthermore, the step of adjusting pheromone concentration based on real-time medical data and the current pheromone matrix according to medical needs includes:

[0042] Based on seasonal disease outbreak data, the release of pheromones in relevant departments and nodes will be temporarily increased.

[0043] Furthermore, the step of constructing a medical 3D network to map medical terms to 3D coordinate nodes and calculating node distances using a co-occurrence matrix includes:

[0044] Determine the co-occurrence relationships among medical terms;

[0045] Count the number of times each pair of medical terms co-occurs in the dataset;

[0046] A co-occurrence matrix was constructed using medical terms as rows and columns, and co-occurrence frequencies as elements.

[0047] Calculate the distance between medical terms based on the co-occurrence matrix;

[0048] Mapping medical terminology to three-dimensional space;

[0049] Each medical term is treated as a node in a three-dimensional space, and the nodes are connected according to the calculated distance to form a three-dimensional network.

[0050] Furthermore, the step of projecting the medical 3D network onto a 2D plane based on each patient's attributes to form an interactive 2D cognitive map includes:

[0051] The projection rules are determined based on the patient's attributes;

[0052] Transform three-dimensional coordinates into two-dimensional coordinates using a projection function;

[0053] For each node in the 3D network, calculate its 2D coordinates according to the projection rules;

[0054] The projected 2D coordinates are used as node positions, and the connection relationships in the 3D network are preserved to construct a 2D network.

[0055] In summary, the intelligent management method for medical and health record information of this invention constructs a three-dimensional medical network, mapping medical terms to three-dimensional coordinate nodes and using a co-occurrence matrix to calculate node distances. This comprehensively integrates medical information, organizing different types of medical terms such as diseases, symptoms, and drugs in a structured manner, while accurately reflecting the degree of correlation between medical terms. Based on patient attributes, the three-dimensional medical network is projected onto a two-dimensional plane, forming an interactive two-dimensional cognitive map. For example, pediatric disease nodes can be automatically aggregated to the left side of the plane, facilitating quick location of relevant disease information for pediatric patients and making it easier to understand and operate. The correlation between different medical terms can be quickly understood based on node distribution and distance. The record classification rules are optimized based on the two-dimensional cognitive map. Reasonable classification rules enable rapid location of categories that may contain the target record during retrieval, reducing the search scope and improving the accuracy of search results while avoiding interference from irrelevant records. Based on the optimal record classification rule set and the two-dimensional cognitive map, the optimal retrieval path is planned, further improving retrieval efficiency. The system can plan the shortest retrieval path on the two-dimensional map based on the location and related information of the target record, quickly locating and obtaining the target record according to the order of record nodes along the path, without having to search through complex record structures. This invention constructs a three-dimensional network to integrate information, reflect terminological relationships, projects it into a two-dimensional cognitive map, and optimizes classification rules and optimal path planning. By following the optimal path, the target file can be accurately found, and irrelevant files can be avoided, greatly improving retrieval efficiency and accuracy.

[0056] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by means of embodiments of the invention. Attached Figure Description

[0057] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0058] Figure 1 This is a flowchart of the intelligent management method for medical and health record information according to Embodiment 1 of the present invention;

[0059] Figure 2 This is a system block diagram of the intelligent management system for medical and health record information according to Embodiment 2 of the present invention. Detailed Implementation

[0060] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Several embodiments of the invention are illustrated in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0061] It should be noted that when a component is said to be "fixed to" another component, it can be directly on the other component or there may be an intervening component. When a component is said to be "connected to" another component, it can be directly connected to the other component or there may be an intervening component. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0062] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0063] Example 1

[0064] Please see Figure 1 This invention proposes an intelligent management method for medical and health record information, which includes steps S101 to S105:

[0065] S101, construct a medical 3D network to map medical terms to 3D coordinate nodes, and calculate node distances through a co-occurrence matrix.

[0066] It's important to note that constructing a three-dimensional medical network to integrate medical information is crucial. The medical field contains a vast amount of information, such as various diseases, symptoms, and medications. This information is scattered and complex, making it difficult to effectively integrate and utilize using traditional management methods. By constructing a three-dimensional medical network and mapping medical terms to three-dimensional coordinate nodes, this scattered information can be organized in a structured form, forming a unified knowledge system that facilitates subsequent analysis and processing.

[0067] Medical terms often involve complex relationships, such as the causal relationship between diseases and symptoms, and the therapeutic relationship between diseases and drugs. By constructing a three-dimensional network, these relationships can be visualized more intuitively, revealing potential medical knowledge and patterns, and providing new ideas and methods for medical research, clinical diagnosis, and treatment.

[0068] When mapping medical terms to three-dimensional coordinate nodes, each medical term is explicitly defined as an independent node. For example, "hypertension," "diabetes," "headache," and "aspirin" are all different nodes. The coordinates of each node in three-dimensional space are then determined. This requires considering various factors, such as the semantic features of the medical term, its clinical importance, and its prevalence. Data analysis and machine learning methods, such as cluster analysis and dimensionality reduction algorithms, can be used to map the features of medical terms to three-dimensional coordinate space.

[0069] Mapping medical terms to 3D coordinate nodes requires considering the semantic similarity of the terms; for example, diseases with similar symptoms may be geographically close in 3D space. Common and serious diseases should also be considered, and these can be placed in more central positions in 3D space to highlight their importance. Furthermore, the position of the nodes in 3D space can be adjusted based on factors such as the incidence rate of the disease in different regions and populations.

[0070] A co-occurrence matrix is ​​constructed by collecting large amounts of medical data, such as medical records and medical literature, and statistically analyzing the frequency of different medical terms appearing in the same record. For example, if "hypertension" and "diabetes" appear simultaneously in a medical record, the values ​​of their corresponding positions in the co-occurrence matrix will increase. Based on the data in the co-occurrence matrix, the distance between nodes is calculated using distance metrics such as Euclidean distance and cosine similarity. Taking Euclidean distance as an example, if two medical terms have a high co-occurrence frequency in the co-occurrence matrix, they will be closer; conversely, if the co-occurrence frequency is low, the distance will be greater. Calculating node distances using the co-occurrence matrix accurately reflects the actual associations between medical terms. For example, hypertension and diabetes often co-occur frequently in clinical practice due to complications, and the calculated distance is relatively close, which aligns with actual medical practice. This distance calculation based on real medical data provides a reliable basis for subsequent analysis and applications, such as disease prediction and drug recommendation.

[0071] In step S102, the 3D network is projected onto a 2D plane based on the patient's age, gender, and other attributes, forming an interactive 2D cognitive map. The structure of the 3D network and the distance between nodes directly affect the effectiveness of the 2D projection and the rationality of the 2D cognitive map. In step S103, the file classification rules are further optimized based on the 2D cognitive map. Therefore, the relationships and distances between medical terms in the 3D network will, to some extent, affect the optimization results of the file classification rules, thereby affecting retrieval efficiency and accuracy.

[0072] Further optionally, the step of constructing a medical 3D network to map medical terms to 3D coordinate nodes and calculating node distances using a co-occurrence matrix includes:

[0073] Determine the co-occurrence relationships among medical terms;

[0074] Count the number of times each pair of medical terms co-occurs in the dataset;

[0075] A co-occurrence matrix was constructed using medical terms as rows and columns, and co-occurrence frequencies as elements.

[0076] Calculate the distance between medical terms based on the co-occurrence matrix;

[0077] Mapping medical terminology to three-dimensional space;

[0078] Each medical term is treated as a node in a three-dimensional space, and the nodes are connected according to the calculated distance to form a three-dimensional network.

[0079] This involves identifying co-occurrence relationships between medical terms, such as diseases and symptoms, or diseases and medications, that appear together in the same medical record. It also involves counting the number of times each pair of medical terms co-occurs in the dataset; for example, counting how many medical records contain "hypertension" and "diabetes."

[0080] Construct a co-occurrence matrix using medical terms as rows and columns, and co-occurrence frequencies as elements. For example, construct an n×n matrix C, where n is the number of medical terms, and C... ij Indicates the co-occurrence frequency of terms i and j.

[0081] The distance between medical terms can be calculated from the co-occurrence matrix using an inverse relationship based on co-occurrence frequency, such as d. ij =1 / (C) ij +1), where d ij This indicates the distance between term i and term j.

[0082] Medical terms can be mapped to three-dimensional space using methods such as multidimensional scaling analysis (MDS). MDS is a method that converts a distance matrix into coordinates and finds the optimal coordinate representation by minimizing the stress function. For example, the 3D coordinate calculation can be performed using Python's sklearn.manifold.MDS function.

[0083] Each medical term is treated as a node in a three-dimensional space, and the nodes are connected according to the calculated distance to form a three-dimensional network.

[0084] S102, based on each patient's attributes, projects the medical 3D network onto a 2D plane to form an interactive 2D cognitive map.

[0085] It should be noted that patients of different ages and genders have different needs and concerns regarding medical information. Projecting a three-dimensional medical network onto a two-dimensional plane based on patient attributes such as age and gender can create a personalized two-dimensional cognitive map. For example, pediatric disease nodes are automatically aggregated on the left side of the plane, making it easier for doctors to quickly locate relevant disease information when diagnosing pediatric patients, thus improving the relevance and efficiency of information acquisition.

[0086] Projecting a 3D network onto a 2D plane presents medical information in a more intuitive and concise way. The 2D cognitive map provides a foundation for optimizing record classification rules. By analyzing the distribution and connections of nodes on the 2D network, potential patterns and associations between medical terms can be discovered, leading to the development of more reasonable record classification rules. During record retrieval, the 2D cognitive map helps plan the optimal search path. The system can quickly determine the search direction and path based on the target record's location on the 2D network, improving retrieval efficiency.

[0087] Further optionally, the step of projecting the medical 3D network onto a 2D plane based on each patient's attributes to form an interactive 2D cognitive map includes:

[0088] The projection rules are determined based on the patient's attributes;

[0089] Transform three-dimensional coordinates into two-dimensional coordinates using a projection function;

[0090] For each node in the 3D network, calculate its 2D coordinates according to the projection rules;

[0091] The projected 2D coordinates are used as node positions, and the connection relationships in the 3D network are preserved to construct a 2D network.

[0092] Understandably, projection rules are determined based on patient attributes. For example, disease nodes related to children are automatically aggregated to the left side of the plane, while those related to the elderly are aggregated to the right side; disease nodes related to women are aggregated to the top of the plane, while those related to men are aggregated to the bottom. Patients of different ages and genders suffer from different diseases. This aggregation method can intuitively display the relationship between diseases and patient attributes, facilitating doctors to quickly locate relevant information during diagnosis and treatment. In addition to age and gender attributes, other patient attributes, such as occupation and region, can be considered to further expand the dimensions of the projection rules, enabling the two-dimensional cognitive map to display richer medical information.

[0093] Projection rules can be adjusted and modified according to actual needs. For example, if a change is found in the distribution of a certain disease in a specific age group or gender, the location and method of aggregation can be adjusted accordingly.

[0094] The transformation of three-dimensional coordinates into two-dimensional coordinates can be achieved through a projection function. For example, for a three-dimensional coordinate (x, y, z), based on the age attribute 'a' and the gender attribute 'g' (g=0 represents female, g=1 represents male), the rule for projecting it to the two-dimensional coordinate (x′, y′) could be: x′=x+k1×(a−a) 平均 ), y′=y+k2×(g−0.5), where k1 and k2 are coefficients controlling the projection intensity. Different coefficient values ​​will result in different degrees of change in the position of the node on the two-dimensional plane, a平均 This is the average age of the patients.

[0095] For each node in the 3D network, its 2D coordinates are calculated according to the projection rules. By traversing all nodes in the 3D network and substituting the 3D coordinates of each node into the projection function, the corresponding 2D coordinates can be obtained.

[0096] By using the projected two-dimensional coordinates as node positions and preserving the connections in the three-dimensional network, a two-dimensional network is constructed to display medical information in a planar graphical form, making it easier to quickly understand the distribution and relationships between nodes.

[0097] S103, optimize the archive classification rules based on the two-dimensional cognitive map to generate the optimal archive classification rule set.

[0098] It should be noted that optimizing record classification rules based on two-dimensional cognitive maps makes record classification more consistent with the inherent logic of medical knowledge and practical application needs. This optimization reduces the ambiguity and error rate of record classification, providing a solid foundation for subsequent retrieval operations and ensuring that records can be accurately and quickly located and retrieved.

[0099] A reasonable set of archival classification rules can make the organization of archives more orderly, reduce the search scope during retrieval, thereby improving the efficiency of retrieval path planning and ensuring that the retrieval operation can more accurately locate the target archive.

[0100] Optionally, the step of optimizing the document classification rules based on the two-dimensional cognitive map to generate the optimal document classification rule set includes:

[0101] Encode all file classification rules into binary strings;

[0102] With retrieval efficiency and misdiagnosis avoidance rate as dual objectives, the application effect of archival classification rules on a two-dimensional cognitive map is simulated. The optimal set of archival classification rules is selected through multiple rounds of iteration to guide archival classification and retrieval.

[0103] Understandably, encoding all possible classification rules into binary strings transforms complex classification rules into simple numerical forms, making comparison, selection, and evolution operations between rules more efficient. Suppose there are 5 classification methods (by department, by symptom, by age, by gender, and by disease type). Each rule can be represented by a 5-bit binary string, such as "10101" representing classification by symptom, not by age, by gender, not by disease type, and by department.

[0104] The optimization aims to improve both retrieval efficiency (path shortening rate) and misdiagnosis avoidance rate (conflict rule detection). Retrieval efficiency reflects the extent to which rules improve the speed of document retrieval, while misdiagnosis avoidance rate reflects the ability of rules to avoid misclassification. These two objectives comprehensively consider the effectiveness and security of document classification rules in practical applications. By simulating the application effect of document classification rules on a two-dimensional cognitive map, multiple rounds of iteration are conducted to select the optimal set of document classification rules to guide document classification and retrieval.

[0105] Further, optionally, after the step of selecting the optimal set of document classification rules through multiple rounds of iteration, the method further includes:

[0106] New case data is injected automatically on a regular basis, triggering the mutation and crossover of the file classification rule set, eliminating inefficient rules, adjusting the file classification rule set, and obtaining the updated optimal file classification rule set.

[0107] Understandably, regularly updating the record classification rule set is crucial to maintaining its timeliness and adaptability, ensuring that the rule set keeps pace with developments and changes in the medical field. Continuous updates to medical knowledge and case information, along with the regular injection of new case data that triggers variations and cross-validations in the rule set, ensure that the rule set always reflects the latest medical information. The regularly automatically injected new case data includes information on new disease types, treatment methods, and symptom presentations.

[0108] The process triggers mutations and crossovers in the archival classification rule set, generating new rule individuals by simulating the mutation and crossover mechanisms in biological evolution. Mutation operations randomly alter certain characteristics of rule individuals, while crossover operations exchange some characteristics between different rule individuals, thus increasing the diversity of the rule set. Based on certain evaluation criteria, such as retrieval efficiency and misdiagnosis avoidance rate, poorly performing rules are selected and eliminated. After adjusting the archival classification rule set through mutation, crossover, and elimination operations, an updated optimal archival classification rule set is obtained.

[0109] Optionally, the step of optimizing the retrieval efficiency and misdiagnosis avoidance rate with dual objectives, simulating the application effect of document classification rules on a two-dimensional cognitive map, and selecting the optimal set of document classification rules through multiple rounds of iteration includes:

[0110] Generate initial rule individuals to form an initial population, where the binary string of each file classification rule represents a rule individual;

[0111] On a two-dimensional cognitive map, simulated archival data is classified according to archival classification rules;

[0112] Record the average path length required to retrieve a specific file, both when and after classifying it using file classification rules.

[0113] The path shortening rate is calculated using the formula: Path shortening rate = (Average path length without rules - Average path length with rules) / Average path length without rules × 100%;

[0114] The number of conflicting rules is counted, and the misdiagnosis avoidance rate is calculated using the formula: Misdiagnosis Avoidance Rate = (1 - Number of Conflicting Rules / Total Number of Rules) × 100%;

[0115] The path shortening rate and misdiagnosis avoidance rate are weighted and summed to obtain the overall fitness of each rule individual;

[0116] Select rule individuals with high overall fitness to enter the next generation, and perform crossover and mutation operations to obtain new rule individuals. Repeat the iteration until the overall fitness converges, then terminate the iteration and select the rule individual with the highest overall fitness as the optimal file classification rule.

[0117] Specifically, all possible file classification rules are encoded to form an initial set of binary string rules, with each binary string representing an individual rule.

[0118] A certain number of initial rule individuals are randomly generated to form the initial population.

[0119] Determine the crossover probability and mutation probability. These two parameters are used to control the evolutionary process of regular individuals. The crossover probability can be set between 0.6 and 0.9, and the mutation probability can be set between 0.01 and 0.1.

[0120] For each individual rule, it is decoded into a specific file classification rule so that it can be simulated and applied on a two-dimensional cognitive map.

[0121] On the two-dimensional cognitive map, the simulated archival data is classified according to the decoded archival classification rules. For example, if the rule represents classification by department and symptom, the archival data is classified according to the combination of department and symptom.

[0122] Records the average path length required to retrieve a specific file, both before and after using file classification rules.

[0123] The path shortening rate is calculated using the formula: Path shortening rate = (Average path length without rules - Average path length with rules) / Average path length without rules × 100%.

[0124] Check for conflicting file classification rules within the individual rules. For example, a rule that requires a certain type of file to be classified both by department and by symptoms unrelated to that department could lead to misdiagnosis.

[0125] The number of conflicting rules is counted, and the misdiagnosis avoidance rate is calculated using the formula: Misdiagnosis avoidance rate = (1 - number of conflicting rules / total number of rules) × 100%.

[0126] The path shortening rate and misdiagnosis avoidance rate are weighted and summed to obtain the overall fitness of each rule. The weights can be adjusted according to actual needs; for example, the retrieval efficiency weight can be 0.6, and the misdiagnosis avoidance rate weight can be 0.4. The formula for overall fitness is: Overall Fitness = Retrieval Efficiency Weight × Path Shortening Rate + Misdiagnosis Avoidance Rate Weight × Misdiagnosis Avoidance Rate.

[0127] Individuals with high overall fitness are selected to enter the next generation. A crossover operation is performed on the selected individuals. Two individuals can be randomly selected, and crossover is performed at a certain position in the binary string according to the crossover probability, generating two new individuals. For example, two individuals "10101" and "11010" will generate "10010" and "11101" after crossover at the third position. A mutation operation is then performed on the new individuals generated by crossover. According to the mutation probability, a random gene position in the individual is flipped. For example, the second position of the individual "10010" will mutate to become "11010". The crossover and mutated individuals are then combined to form a new population.

[0128] Repeat the fitness evaluation, selection, crossover, and mutation operations until the maximum number of iterations is reached or the fitness converges. After the iterations are complete, select the rule with the highest overall fitness as the optimal archival classification rule.

[0129] S104, based on the optimal set of document classification rules and a two-dimensional cognitive map, plans the optimal retrieval path.

[0130] It should be noted that the optimal set of record classification rules determines the organization of records and provides a framework for planning retrieval paths. The two-dimensional cognitive map provides the spatial relationships between record nodes. Combined with the set of record classification rules, it is possible to make full use of the organizational structure of records and the relationships between medical terms to formulate efficient retrieval strategies, so as to quickly and accurately find target records during the record retrieval process and reduce retrieval time and resource consumption.

[0131] Further, optionally, the step of planning the optimal retrieval path based on the optimal set of document classification rules and the two-dimensional cognitive map includes:

[0132] Determine the file association strength value;

[0133] The strength of file associations is converted into pheromone concentration, and a pheromone matrix is ​​constructed.

[0134] Based on the pheromone matrix and the optimal set of file classification rules, the system simulates ant colonies releasing pheromones on a two-dimensional cognitive map to perform retrieval, and plans the optimal retrieval path using a probabilistic selection formula.

[0135] Understandably, determining the file association strength value provides a basis for constructing the pheromone matrix. The file association strength reflects the closeness between files, accurately measures the relationship between files, and helps to construct a reasonable pheromone matrix, thereby guiding the ant colony algorithm to plan a better retrieval path.

[0136] The strength of association is calculated by considering multiple factors, such as the similarity of document content, citation relationships, and business logic connections. For text-based documents, the cosine similarity algorithm can be used to calculate the similarity as part of the association strength. For example, for documents related to diabetes and retinopathy, the association strength is determined to be 0.8 based on their association with medical knowledge.

[0137] The strength of the association between archives is converted into pheromone concentration, forming a pheromone matrix. The pheromone matrix reflects the attraction and reachability between archive nodes, providing key environmental information for ant colony algorithms to simulate the retrieval behavior of ants on a two-dimensional cognitive map, and influencing the ants' path selection.

[0138] Further optionally, the step of simulating ant colonies releasing pheromones on a two-dimensional cognitive map for retrieval based on the pheromone matrix and the optimal file classification rule set, and planning the optimal retrieval path using a probabilistic selection formula, includes:

[0139] For each ant in the colony, starting from the initial file node of the two-dimensional cognitive map, the next file node to be visited is selected according to the probability selection formula, and the retrieval path from the initial file to the target file is recorded.

[0140] After all ants have completed one path construction, the pheromone matrix is ​​evaporated.

[0141] For each path constructed by an ant, pheromones are released based on the quality of the path.

[0142] Update the pheromone matrix and add the pheromones released by all ants to the corresponding edges;

[0143] Repeat the iteration until the pheromone matrix tends to stabilize. After the iteration ends, select the path with the highest pheromone concentration as the optimal retrieval path based on the current pheromone matrix and the probability selection formula.

[0144] It is understood that this embodiment simulates the collective behavior of ants through the ant colony algorithm, which can find a better retrieval path in a complex two-dimensional cognitive map based on the pheromone matrix and the optimal archive classification rule set, thereby improving the efficiency of archive retrieval.

[0145] Further, optionally, the step of planning the optimal retrieval path using the probabilistic selection formula may further include:

[0146] Based on real-time medical data and the current pheromone matrix, the pheromone concentration is adjusted according to medical needs to obtain an adjusted pheromone matrix, making the retrieval path more in line with current medical needs.

[0147] Understandably, adjusting pheromone concentrations based on real-time medical data makes search paths more aligned with current medical needs, improving the relevance and usability of searches. Real-time medical data reflects the current medical situation and needs; adjusting pheromone concentrations allows search paths to adapt to these changes, enabling timely adjustments to search strategies based on actual circumstances and improving search effectiveness.

[0148] Further optionally, the step of adjusting the pheromone concentration based on real-time medical data and the current pheromone matrix according to medical needs includes:

[0149] Based on seasonal disease outbreak data, the release of pheromones in relevant departments and nodes will be temporarily increased.

[0150] Understandably, based on real-time medical information such as seasonal disease outbreak data, the pheromone release of relevant department nodes can be temporarily increased. For example, during flu season, the pheromone release of respiratory department nodes can be increased to make search paths more inclined to access respiratory-related files.

[0151] S105: Perform file retrieval operations on the two-dimensional cognitive map using the optimal retrieval path, and quickly locate and obtain the target file based on the order of file nodes on the path.

[0152] In summary, the intelligent management system for medical and health record information of this invention constructs a three-dimensional medical network, mapping medical terms to three-dimensional coordinate nodes. It uses a co-occurrence matrix to calculate node distances, comprehensively integrating medical information and organizing different types of medical terms such as diseases, symptoms, and medications in a structured manner. Simultaneously, it accurately reflects the degree of correlation between medical terms. Based on patient attributes, the three-dimensional medical network is projected onto a two-dimensional plane, forming an interactive two-dimensional cognitive map. For example, pediatric disease nodes can be automatically aggregated to the left side of the plane, facilitating rapid location of relevant disease information for pediatric patients and making it easier to understand and operate. The system allows for quick understanding of the correlation between different medical terms based on node distribution and distance. Based on the two-dimensional cognitive map, the record classification rules are optimized. Reasonable classification rules enable rapid location of categories that may contain the target record during retrieval, reducing the search scope and improving the accuracy of search results while avoiding interference from irrelevant records. Based on the optimal record classification rule set and the two-dimensional cognitive map, the optimal search path is planned, further improving search efficiency. The system can plan the shortest search path on the two-dimensional map based on the location and relevant information of the target record, quickly locating and retrieving the target record according to the order of record nodes along the path, without having to search through complex record structures. This invention constructs a three-dimensional network to integrate information, reflect terminological relationships, projects it into a two-dimensional cognitive map, and optimizes classification rules and optimal path planning. By following the optimal path, the target file can be accurately found, and irrelevant files can be avoided, greatly improving retrieval efficiency and accuracy.

[0153] Example 2

[0154] Please see Figure 2 The present invention proposes an intelligent management system for medical and health record information, which includes:

[0155] 3D Network Module: Used to construct a medical 3D network to map medical terms to 3D coordinate nodes and calculate node distances through a co-occurrence matrix;

[0156] Two-dimensional map module: used to project the medical three-dimensional network onto a two-dimensional plane based on each patient's attributes, forming an interactive two-dimensional cognitive map;

[0157] Classification rules module: used to optimize the classification rules of archives based on the two-dimensional cognitive map and generate the optimal set of archive classification rules;

[0158] Search path module: Used to plan the optimal search path based on the optimal set of document classification rules and two-dimensional cognitive map;

[0159] Document retrieval module: Used to perform document retrieval operations on a two-dimensional cognitive map using the optimal retrieval path, and quickly locate and retrieve target documents based on the order of document nodes on the path.

[0160] Further optionally, the classification rule module is also used for:

[0161] Encode all file classification rules into binary strings;

[0162] With retrieval efficiency and misdiagnosis avoidance rate as dual objectives, the application effect of archive classification rules on a two-dimensional cognitive map is simulated. The optimal set of archive classification rules is selected through multiple rounds of iteration to guide archive classification and retrieval.

[0163] New case data is injected automatically on a regular basis, triggering the mutation and crossover of the file classification rule set, eliminating inefficient rules, adjusting the file classification rule set, and obtaining the updated optimal file classification rule set.

[0164] Further optionally, the classification rule module is also used for:

[0165] Generate initial rule individuals to form an initial population, where the binary string of each file classification rule represents a rule individual;

[0166] On a two-dimensional cognitive map, simulated archival data is classified according to archival classification rules;

[0167] Record the average path length required to retrieve a specific file, both when and after classifying it using file classification rules.

[0168] The path shortening rate is calculated using the formula: Path shortening rate = (Average path length without rules - Average path length with rules) / Average path length without rules × 100%;

[0169] The number of conflicting rules is counted, and the misdiagnosis avoidance rate is calculated using the formula: Misdiagnosis Avoidance Rate = (1 - Number of Conflicting Rules / Total Number of Rules) × 100%;

[0170] The path shortening rate and misdiagnosis avoidance rate are weighted and summed to obtain the overall fitness of each rule individual;

[0171] Select rule individuals with high overall fitness to enter the next generation, and perform crossover and mutation operations to obtain new rule individuals. Repeat the iteration until the overall fitness converges, then terminate the iteration and select the rule individual with the highest overall fitness as the optimal file classification rule.

[0172] Further optionally, the retrieval path module is also used for:

[0173] Determine the file association strength value;

[0174] The strength of file associations is converted into pheromone concentration, and a pheromone matrix is ​​constructed.

[0175] Based on the pheromone matrix and the optimal set of file classification rules, the ant colony releases pheromones on a two-dimensional cognitive map to perform retrieval, and plans the optimal retrieval path through a probabilistic selection formula.

[0176] Further optionally, the retrieval path module is also used for:

[0177] For each ant in the colony, starting from the initial file node of the two-dimensional cognitive map, the next file node to be visited is selected according to the probability selection formula, and the retrieval path from the initial file to the target file is recorded.

[0178] After all ants have completed one path construction, the pheromone matrix is ​​evaporated.

[0179] For each path constructed by an ant, pheromones are released based on the quality of the path.

[0180] Update the pheromone matrix and add the pheromones released by all ants to the corresponding edges;

[0181] Repeat the iteration until the pheromone matrix tends to stabilize. After the iteration ends, select the path with the highest pheromone concentration as the optimal retrieval path based on the current pheromone matrix and the probability selection formula.

[0182] Further optionally, the retrieval path module is also used for:

[0183] Based on real-time medical data and the current pheromone matrix, the pheromone concentration is adjusted according to medical needs to obtain an adjusted pheromone matrix, making the retrieval path more in line with current medical needs.

[0184] Further optionally, the retrieval path module is also used for:

[0185] Based on seasonal disease outbreak data, the release of pheromones in relevant departments and nodes will be temporarily increased.

[0186] Further optionally, the three-dimensional network module is also used for:

[0187] Determine the co-occurrence relationships among medical terms;

[0188] Count the number of times each pair of medical terms co-occurs in the dataset;

[0189] A co-occurrence matrix was constructed using medical terms as rows and columns, and co-occurrence frequencies as elements.

[0190] Calculate the distance between medical terms based on the co-occurrence matrix;

[0191] Mapping medical terminology to three-dimensional space;

[0192] Each medical term is treated as a node in a three-dimensional space, and the nodes are connected according to the calculated distance to form a three-dimensional network.

[0193] Further optionally, the two-dimensional map module is also used for:

[0194] The projection rules are determined based on the patient's attributes;

[0195] Transform three-dimensional coordinates into two-dimensional coordinates using a projection function;

[0196] For each node in the 3D network, calculate its 2D coordinates according to the projection rules;

[0197] The projected 2D coordinates are used as node positions, and the connection relationships in the 3D network are preserved to construct a 2D network.

[0198] In summary, the intelligent management system for medical and health record information of this invention constructs a medical 3D network through a 3D network module, mapping medical terms to 3D coordinate nodes and calculating node distances using a co-occurrence matrix. This comprehensively integrates medical information, organizing different types of medical terms such as diseases, symptoms, and medications in a structured manner, while accurately reflecting the degree of correlation between medical terms. Furthermore, the 2D map module projects the medical 3D network onto a 2D plane based on patient attributes, forming an interactive 2D cognitive map. For example, it can automatically aggregate pediatric disease nodes to the left side of the plane, facilitating quick location of relevant disease information for pediatric patients and making the system easier to understand and operate. The system can also be configured based on node distribution and distance. This invention enables rapid understanding of the relationships between different medical terms. A classification rule module optimizes file classification rules based on a two-dimensional cognitive map. Reasonable classification rules allow for quick location of categories likely containing the target file during retrieval, reducing the search scope and improving the accuracy of search results while avoiding interference from irrelevant files. A retrieval path module plans the optimal retrieval path based on the optimal set of file classification rules and the two-dimensional cognitive map, further improving retrieval efficiency. The system can plan the shortest retrieval path on the two-dimensional map based on the location and relevant information of the target file. The file retrieval module quickly locates and retrieves the target file according to the order of file nodes along the path, eliminating the need to search through complex file structures. This invention constructs a three-dimensional network to integrate information, reflect terminological relationships, projects it into a two-dimensional cognitive map, and optimizes classification rules and plans the optimal path. Retrieving the target file accurately using the optimal path avoids irrelevant files, greatly improving retrieval efficiency and accuracy.

[0199] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A medical and health record information intelligent management system, characterized in that, The system includes: 3D Network Module: Used to construct a medical 3D network to map medical terms to 3D coordinate nodes and calculate node distances through a co-occurrence matrix; Two-dimensional map module: used to project the medical three-dimensional network onto a two-dimensional plane based on each patient's attributes, forming an interactive two-dimensional cognitive map; Classification rules module: used to optimize the classification rules of archives based on the two-dimensional cognitive map and generate the optimal set of archive classification rules; Search path module: Used to plan the optimal search path based on the optimal set of document classification rules and two-dimensional cognitive map; Document retrieval module: Used to perform document retrieval operations on a two-dimensional cognitive map using the optimal retrieval path, and quickly locate and retrieve target documents based on the order of document nodes on the path.

2. A method for intelligent management of medical and health record information, used to implement the intelligent management system for medical and health record information as described in claim 1, characterized in that, The method includes: A medical 3D network is constructed to map medical terms to 3D coordinate nodes, and the node distance is calculated through a co-occurrence matrix; Based on each patient's attributes, the medical 3D network is projected onto a 2D plane to form an interactive 2D cognitive map; The document classification rules are optimized based on a two-dimensional cognitive map to generate the optimal document classification rule set; Based on the optimal set of archive classification rules and a two-dimensional cognitive map, the optimal retrieval path is planned; The optimal search path is used to perform file retrieval on a two-dimensional cognitive map, and the target file is quickly located and retrieved based on the order of file nodes on the path.

3. The intelligent management method for medical and health record information according to claim 2, characterized in that, The steps for optimizing the document classification rules based on a two-dimensional cognitive map to generate the optimal document classification rule set include: Encode all file classification rules into binary strings; With retrieval efficiency and misdiagnosis avoidance rate as dual objectives, the application effect of archive classification rules on a two-dimensional cognitive map is simulated. The optimal set of archive classification rules is selected through multiple rounds of iteration to guide archive classification and retrieval. New case data is injected automatically on a regular basis, triggering the mutation and crossover of the file classification rule set, eliminating inefficient rules, adjusting the file classification rule set, and obtaining the updated optimal file classification rule set.

4. The intelligent management method for medical and health record information according to claim 3, characterized in that, The steps of optimizing the system with the dual objectives of retrieval efficiency and misdiagnosis avoidance rate, simulating the application effect of document classification rules on a two-dimensional cognitive map, and selecting the optimal set of document classification rules through multiple rounds of iteration include: Generate initial rule individuals to form an initial population, where the binary string of each file classification rule represents a rule individual; On a two-dimensional cognitive map, simulated archival data is classified according to archival classification rules; Record the average path length required to retrieve a specific file, both when and after classifying it using file classification rules. The path shortening rate is calculated using the formula: Path shortening rate = (Average path length without rules - Average path length with rules) / Average path length without rules × 100%; The number of conflicting rules is counted, and the misdiagnosis avoidance rate is calculated using the formula: Misdiagnosis Avoidance Rate = (1 - Number of Conflicting Rules / Total Number of Rules) × 100%; The path shortening rate and misdiagnosis avoidance rate are weighted and summed to obtain the overall fitness of each rule individual; Select rule individuals with high overall fitness to enter the next generation, and perform crossover and mutation operations to obtain new rule individuals. Repeat the iteration until the overall fitness converges, then terminate the iteration and select the rule individual with the highest overall fitness as the optimal file classification rule.

5. The intelligent management method for medical and health record information according to claim 2, characterized in that, The steps for planning the optimal retrieval path based on the optimal set of document classification rules and a two-dimensional cognitive map include: Determine the file association strength value; The strength of file associations is converted into pheromone concentration, and a pheromone matrix is ​​constructed. Based on the pheromone matrix and the optimal set of file classification rules, the ant colony releases pheromones on a two-dimensional cognitive map to perform retrieval, and plans the optimal retrieval path through a probabilistic selection formula.

6. The intelligent management method for medical and health record information according to claim 5, characterized in that, The steps of simulating ant colonies releasing pheromones on a two-dimensional cognitive map for retrieval based on the pheromone matrix and the optimal file classification rule set, and planning the optimal retrieval path using a probabilistic selection formula, include: For each ant in the colony, starting from the initial file node of the two-dimensional cognitive map, the next file node to be visited is selected according to the probability selection formula, and the retrieval path from the initial file to the target file is recorded. After all ants have completed one path construction, the pheromone matrix is ​​evaporated. For each path constructed by an ant, pheromones are released based on the quality of the path. Update the pheromone matrix and add the pheromones released by all ants to the corresponding edges; Repeat the iteration until the pheromone matrix tends to stabilize. After the iteration ends, select the path with the highest pheromone concentration as the optimal retrieval path based on the current pheromone matrix and the probability selection formula.

7. The intelligent management method for medical and health record information according to claim 5, characterized in that, The step of planning the optimal retrieval path using the probabilistic selection formula further includes: Based on real-time medical data and the current pheromone matrix, the pheromone concentration is adjusted according to medical needs to obtain an adjusted pheromone matrix, making the retrieval path more in line with current medical needs.

8. The intelligent management method for medical and health record information according to claim 7, characterized in that, The step of adjusting pheromone concentration based on real-time medical data and the current pheromone matrix according to medical needs includes: Based on seasonal disease outbreak data, the release of pheromones in relevant departments and nodes will be temporarily increased.

9. The intelligent management method for medical and health record information according to claim 1, characterized in that, The steps of constructing a medical 3D network to map medical terms to 3D coordinate nodes and calculating node distances using a co-occurrence matrix include: Determine the co-occurrence relationships among medical terms; Count the number of times each pair of medical terms co-occurs in the dataset; A co-occurrence matrix was constructed using medical terms as rows and columns, and co-occurrence frequencies as elements. Calculate the distance between medical terms based on the co-occurrence matrix; Mapping medical terminology to three-dimensional space; Each medical term is treated as a node in a three-dimensional space, and the nodes are connected according to the calculated distance to form a three-dimensional network.

10. The intelligent management method for medical and health record information according to claim 1, characterized in that, The step of projecting a medical 3D network onto a 2D plane based on each patient's attributes to form an interactive 2D cognitive map includes: The projection rules are determined based on the patient's attributes; Transform three-dimensional coordinates into two-dimensional coordinates using a projection function; For each node in the 3D network, calculate its 2D coordinates according to the projection rules; The projected 2D coordinates are used as node positions, and the connection relationships in the 3D network are preserved to construct a 2D network.

Images