Regional medical data analysis processing method and system based on big data technology

Abstract

The invention provides a regional medical data analysis processing method and system based on a big data technology, and the method comprises the following steps: collecting operation data of all medical institutions in a region, constructing a service pressure prediction model, and calculating a real-time load index; and mapping the load index to a geographic space to form a dynamic distribution diagram. When a user requests a particular medical institution service, it is automatically determined whether the institution exceeds a congestion threshold. And if the target mechanism is overloaded, user demands are converted into standardized demand description, and the most suitable alternative mechanism is screened out through an intelligent matching algorithm. And finally, based on the load comparison condition of the target mechanism and the alternative mechanism, generating a personalized shunting guide suggestion and pushing the suggestion to the user. The method can effectively improve the utilization efficiency of regional medical resources.

Description

Technical Field

[0001] The present invention belongs to the field of big data processing technology, and specifically relates to a regional medical data analysis and processing method and system based on big data technology. Background Art

[0002] With the aging population and rising health awareness, demand for medical services is rapidly increasing. However, high-quality medical resources are relatively scarce and unevenly distributed, leading to chronic overload in large tertiary hospitals and underutilization of primary healthcare facilities. Traditional medical service systems primarily rely on simple appointment queuing mechanisms, lacking intelligent resource allocation capabilities. Despite the rapid development of internet healthcare, while services such as online appointment booking and remote consultation have become more widespread, they still fail to address the fundamental issue of unbalanced resource allocation. This imbalance not only reduces overall medical service efficiency but also increases patient wait times and medical costs, compromising the quality of care. Summary of the Invention

[0003] The present invention provides a regional medical data analysis and processing method and system based on big data technology to solve the above technical problems.

[0004] In a first aspect, the present invention provides a method for analyzing and processing regional medical data based on big data technology, the method comprising the following steps:

[0005] Collect multi-dimensional dynamic data of each medical institution in the region, build a composite pressure index model based on the multi-dimensional dynamic data to predict the service pressure trend of each medical institution, and calculate the real-time composite pressure index of each medical institution;

[0006] Map the real-time composite pressure index of each medical institution to geographic space to construct a dynamic pressure potential field representing the degree of medical service congestion in the region;

[0007] When receiving a service request from a user for a target medical institution, determining whether the real-time composite pressure index of the target medical institution exceeds a preset congestion threshold;

[0008] If the congestion threshold is exceeded, the service request is converted into a patient demand vector. Based on the spatiotemporal distribution characteristics of the dynamic pressure potential field, a hospital service vector with a dynamic attractiveness dimension is constructed for other medical institutions in the area. By calculating the similarity between the patient demand vector and the service vectors of each hospital, one or more alternative medical institutions are selected.

[0009] A diversion guidance strategy is generated based on the pressure characteristics of the target medical institution and the alternative medical institutions in the dynamic pressure potential energy field, and the diversion guidance strategy is pushed to the user terminal.

[0010] Optionally, the step of constructing a composite pressure index model for predicting the service pressure trend of each medical institution based on multi-dimensional dynamic data and calculating the real-time composite pressure index of each medical institution includes the following steps:

[0011] The static resource data, quasi-real-time business data, and real-time queue data in the multi-dimensional dynamic data are fused into the current pressure snapshot value through a weighted fusion function;

[0012] Use the historical current pressure snapshot value sequence as training data to train the pre-deployed time series prediction model;

[0013] Utilize the time series prediction model and output the future pressure forecast value based on the current pressure snapshot value;

[0014] The current pressure snapshot value and the future pressure forecast value are dynamically weighted and combined to generate a real-time composite pressure index.

[0015] Optionally, the time series prediction model is a long short-term memory network model.

[0016] Optionally, mapping the real-time composite pressure index of each medical institution to a geographic space to construct a dynamic pressure potential energy field representing the degree of medical service congestion in the region includes the following steps:

[0017] Each medical institution is abstracted as a virtual mass point in the geographic coordinate system, and the real-time composite pressure index of the medical institution is used as the mass point value of the virtual mass point;

[0018] Based on the geographical locations and particle values ​​of all virtual particles, a spatial interpolation algorithm is used to calculate the pressure potential energy value of each preset grid cell in the geographic coordinate system;

[0019] The pressure potential energy values ​​of all grid cells are aggregated to generate a dynamic pressure potential energy field that covers the entire area and changes continuously.

[0020] Optionally, the process of constructing a hospital service vector including a dynamic attraction dimension for other medical institutions in the region based on the spatiotemporal distribution characteristics of the dynamic pressure potential energy field, and screening one or more candidate medical institutions by calculating the similarity between the patient demand vector and the service vector of each hospital, comprises the following steps:

[0021] Extract the spatiotemporal distribution characteristics of the dynamic pressure potential energy field. These characteristics include the pressure potential energy value and pressure potential energy gradient at each medical institution. The pressure potential energy gradient characterizes the rate and direction of pressure change at the corresponding location in the dynamic pressure potential energy field and is used to quantify whether the service pressure of a medical institution is easing or increasing.

[0022] Based on the keywords and department information in the service request, a word embedding model is used to generate a patient demand vector;

[0023] For medical institutions in the region, a hospital service vector is constructed based on the service capability characteristics, the inverse of the pressure potential energy value, and the pressure potential energy gradient of the medical institutions, which includes the static service dimension, the dynamic attraction dimension, and the pressure trend dimension.

[0024] The cosine similarity algorithm is used to calculate the similarity score between the patient demand vector and each hospital service vector;

[0025] Arrange in descending order according to the similarity score, and select one or more medical institutions with scores higher than the preset matching threshold as alternative medical institutions.

[0026] Optionally, generating a diversion guidance strategy based on the pressure characteristics of the target medical institution and the alternative medical institutions in the dynamic pressure potential energy field, and pushing the diversion guidance strategy to the user terminal of the user includes the following steps:

[0027] Extracting pressure potential energy values ​​of target medical institutions and alternative medical institutions from the dynamic pressure potential energy field;

[0028] Calculate the user's estimated waiting time at the target medical institution and alternative medical institutions based on the pressure potential energy value;

[0029] Combining pressure potential energy value and estimated queue time to generate a structured diversion and guidance strategy;

[0030] The diversion guidance strategy is presented in the form of a pop-up window on the interface for confirming the reservation on the user's user terminal.

[0031] Optionally, the method further comprises the following steps:

[0032] When a user rejects the diversion and guidance strategy, the user's rejection behavior is defined as a rejection event;

[0033] Capturing and associating contextual parameters when a rejection event occurs, the contextual parameters at least include the patient demand vector, the service capability characteristics of the target medical institution and the alternative medical institutions, the difference in pressure potential energy values ​​between the target medical institution and the alternative medical institutions, and the difference in geographical distance;

[0034] Aggregate multiple rejection events and associated contextual parameters to form a decision friction dataset;

[0035] A decision friction map is constructed based on the decision friction dataset, and the decision friction map is used to dynamically adjust the diversion and guidance strategy.

[0036] Optionally, constructing a decision friction map based on the decision friction dataset includes the following steps:

[0037] An unsupervised learning clustering algorithm is used to perform cluster analysis on the context parameters in the decision friction dataset and identify various rejection patterns.

[0038] Each identified rejection pattern is defined as a decision friction node;

[0039] Node association edges are constructed by analyzing the association strength between patient demand vectors and decision friction nodes;

[0040] All decision friction nodes and node-associated edges are integrated to form a decision friction map that represents the internal reasons why user groups refuse to divert.

[0041] Optionally, the dynamically adjusting the diversion and guidance strategy using the decision friction map includes the following steps:

[0042] When screening alternative medical institutions for a user's new service request, the decision friction graph is searched for the target decision friction node that is most strongly associated with the user's patient demand vector;

[0043] Identify the rejection patterns represented by the target decision friction nodes;

[0044] Dynamic adjustment rules are generated based on the rejection pattern, and are applied to the step of screening alternative medical institutions to dynamically adjust the weight parameters when calculating the similarity between the patient demand vector and the service vectors of each hospital.

[0045] In the second aspect, the present invention also provides a regional medical data analysis and processing system based on big data technology, comprising a memory, a processor, and a computer program stored on the memory and runnable on the processor. When the processor executes the computer program, it implements the regional medical data analysis and processing method based on big data technology as described in the first aspect.

[0046] The beneficial effects of the present invention are:

[0047] By establishing an intelligent dynamic monitoring mechanism, the present invention can grasp the service status of each medical institution in the region in real time, completely changing the traditional static management model and realizing a fundamental transformation from passive response to active deployment. The present invention has powerful predictive analysis capabilities, can accurately identify the congestion trends and distribution characteristics of medical services, provide a reliable basis for scientific decision-making, and greatly improve the accuracy and foresight of resource allocation. By constructing an intelligent matching mechanism, the present invention realizes the precise connection between patient needs and medical resources, avoids the waste of resources caused by blind medical treatment, and significantly improves the overall service efficiency. The present invention also has self-learning and optimization capabilities, and can continuously adjust and improve service strategies according to user feedback to form a virtuous circle. The application of the present invention will significantly alleviate the problem of uneven distribution of medical resources, improve the accessibility of medical services, reduce the cost of medical treatment for patients, and provide strong technical support for building a more equitable and efficient medical service system. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a flowchart of a regional medical data analysis and processing method based on big data technology in one embodiment of the present application.

[0049] Figure 2 This is a process framework diagram for generating a real-time composite pressure index using a time series prediction model in one embodiment of the present application.

[0050] Figure 3 A schematic diagram of the process of constructing a decision friction map in one embodiment of the present application. DETAILED DESCRIPTION

[0051] The following will be combined with the accompanying drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the embodiments described are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of this application.

[0052] The terms "first," "second," and the like in the specification and claims of this application are used to distinguish similar objects, and are not used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of this application can be implemented in an order other than that illustrated or described herein, and that the objects distinguished by "first," "second," and the like are generally of the same type, and do not limit the number of objects; for example, the first object can be one or more. In addition, the term "and / or" in the specification and claims refers to at least one of the connected objects, and the character " / " generally indicates that the objects connected are in an "or" relationship.

[0053] Figure 1FIG. 1 is a flow chart of a method for analyzing and processing regional medical data based on big data technology in one embodiment. Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. In addition, Figure 1 At least part of the steps in the above process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but may be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but may be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps. Figure 1 As shown, the regional medical data analysis and processing method based on big data technology disclosed in the present invention specifically includes the following steps:

[0054] S101. Collect multi-dimensional dynamic data of each medical institution in the region, build a composite pressure index model for predicting the service pressure trend of each medical institution based on the multi-dimensional dynamic data, and calculate the real-time composite pressure index of each medical institution.

[0055] Among them, in order to accurately quantify and predict the service pressure of each hospital, it is necessary to comprehensively collect multi-dimensional dynamic data of all medical institutions in the region. These data include relatively fixed static resource data such as the number of beds and the number of doctors, as well as quasi-real-time business data such as the number of registrations and visits, as well as the real-time number of people queuing in each department. Through a weighted fusion function, these data of different natures can be integrated into a "current pressure snapshot value" that can reflect the current load status of the hospital. After that, a pre-built time series prediction model, such as a long short-term memory network (LSTM), is trained using a continuous sequence of pressure snapshot values ​​in history to enable it to learn the inherent laws of pressure changes. Finally, by dynamically weighting the "future pressure prediction value" output by the model based on the current snapshot value and the "current pressure snapshot value", a more forward-looking real-time composite pressure index can be calculated. The index (in is the current pressure snapshot value, is the predicted value of future pressure, is a dynamic weight coefficient), which has the effect of generating an accurate pressure indicator that reflects both reality and predicts the future, laying a solid data foundation for subsequent regional analysis and scheduling.

[0056] S102. Map the real-time composite pressure index of each medical institution to the geographic space, and construct a dynamic pressure potential field that represents the degree of medical service congestion in the region.

[0057] In order to transform the isolated pressure index of each medical institution into a regional, visual congestion situation, it is necessary to map these indices to geographic space and construct a dynamic pressure potential field. In this process, each medical institution is abstracted as a virtual particle with a specific value in the geographic information system (GIS). Its geographic coordinates are its actual location, and its particle value is the previously calculated real-time composite pressure index. Subsequently, a spatial interpolation algorithm, such as the inverse distance weighted method (IDW), is used to calculate the pressure potential energy value of each preset grid cell in the geographic coordinate system of the entire region. The core idea of ​​this algorithm is that the potential energy value of any unknown point is the weighted average of the values ​​of all known particles (i.e. medical institutions) around it, and the weight is inversely proportional to the distance. Specifically, it can be calculated by the formula Calculated, where V is the pressure potential energy value of the grid cell, is the distance from the cell to the i-th medical institution, and k is the power exponent of the distance. By calculating and aggregating the potential energy values ​​of all grid cells, a continuously changing dynamic heat map covering the entire area is generated, visually demonstrating the congestion level and distribution of medical service resources.

[0058] S103. When receiving a service request from a user for a target medical institution, determine whether the real-time composite pressure index of the target medical institution exceeds a preset congestion threshold.

[0059] When the system receives a service request from a user for a target medical institution (e.g. online registration), a key judgment step will be triggered. The system will immediately obtain the real-time composite pressure index of the target medical institution and compare it with a pre-set congestion threshold. This threshold is scientifically set based on historical data and management experience, representing the critical point where the hospital's service capacity is close to saturation. The implementation principle of this judgment step is very direct, which is a simple conditional judgment: If , the subsequent diversion and guidance process is initiated. This step proactively identifies service requests that may result in long wait times and a poor medical experience for users, ensuring that the diversion and guidance mechanism is activated only when intervention is truly necessary. This avoids unnecessary interruptions to regular medical users, improving system efficiency while ensuring a smooth user experience.

[0060] S104. If the congestion threshold is exceeded, the service request is converted into a patient demand vector, and based on the spatiotemporal distribution characteristics of the dynamic pressure potential field, a hospital service vector containing a dynamic attraction dimension is constructed for other medical institutions in the area. By calculating the similarity between the patient demand vector and the service vectors of each hospital, one or more alternative medical institutions are screened out.

[0061] If the pressure index of the target medical institution exceeds the congestion threshold, the system will start the intelligent matching process to find alternative institutions. First, the user's service request (such as "children have a fever, see an internal medicine doctor") will be converted into a patient demand vector containing specific medical needs using word embedding technology (such as Word2Vec). At the same time, the system will build a multi-dimensional hospital service vector for other medical institutions in the region. This vector not only includes static service dimensions such as its specialty capabilities, but also innovatively incorporates dynamic attraction dimensions (for example, it is inversely proportional to the pressure potential energy value V, the less congested, the higher the attraction) and pressure trend dimensions (i.e., the gradient of the pressure potential energy field, indicating whether the pressure is easing or increasing). Subsequently, by calculating the cosine similarity between the patient demand vector and the service vector of each hospital , we can quantify the degree of match between each candidate hospital and the patient's needs. Ultimately, hospitals with the highest similarity scores that exceed a preset matching threshold are selected as candidates. This step achieves precise matching beyond geographic location, taking into account medical expertise, real-time congestion conditions, and future trends.

[0062] S105. Generate a diversion guidance strategy based on the pressure characteristics of the target medical institution and the alternative medical institutions in the dynamic pressure potential energy field, and push the diversion guidance strategy to the user terminal.

[0063] After selecting suitable alternative medical institutions, the last step is to generate and push a persuasive diversion and guidance strategy. The core of this strategy is to transform the abstract "pressure" into concrete information that users can intuitively perceive based on the pressure characteristics in the dynamic pressure potential field. The system extracts the pressure potential energy value V of the target medical institution originally planned by the user and the locations of all alternative medical institutions, and uses a preset function to calculate the pressure potential energy value V. Convert them into estimated waiting time . Then, this information, together with the distance and specialty of the hospital, is integrated into a structured comparison list. For example, on the interface where the user is about to confirm the appointment, it is clearly presented in the form of a pop-up window: "The hospital A you selected is currently extremely busy, and the estimated waiting time is more than 90 minutes. We recommend you Hospital B, which has a high degree of specialist matching, an estimated waiting time of only 20 minutes, and is only a 10-minute drive away from you." The effect of this approach is that by providing quantitative, transparent and easy-to-understand comparison information at key decision-making nodes, it greatly increases the possibility of users accepting diversion suggestions, thereby effectively guiding customer flow without forcing users, achieving a balanced distribution of regional medical resources, and ultimately improving the overall medical experience of patients.

[0064] In one embodiment, constructing a composite pressure index model for predicting the service pressure trend of each medical institution based on multi-dimensional dynamic data and calculating the real-time composite pressure index of each medical institution includes the following steps:

[0065] The static resource data, quasi-real-time business data, and real-time queue data in the multi-dimensional dynamic data are fused into the current pressure snapshot value through a weighted fusion function;

[0066] Use the historical current pressure snapshot value sequence as training data to train the pre-deployed time series prediction model;

[0067] Utilize the time series prediction model and output the future pressure forecast value based on the current pressure snapshot value;

[0068] The current pressure snapshot value and the future pressure forecast value are dynamically weighted and combined to generate a real-time composite pressure index.

[0069] In this embodiment, in order to accurately reflect the immediate service load of the medical institution, it is first necessary to integrate the collected multi-dimensional dynamic data into a single current pressure snapshot value through a weighted fusion function. These data include static resource data representing service potential (such as doctors, number of beds), quasi-real-time business data reflecting the current business volume (such as registration, number of patients), and real-time queuing data reflecting congestion bottlenecks. Since these data units and dimensions are different, they must first be normalized and uniformly mapped to the range of 0 to 1. Subsequently, by assigning different weights to different data items to reflect the differences in their impact on service pressure, the current pressure snapshot value is finally calculated. The calculation method can be expressed as ,in is the normalized data value of the i-th item, The effect of this step is to refine complex and scattered multi-source information into a standardized, horizontally comparable real-time pressure indicator, providing high-quality input for subsequent trend forecasting.

[0070] After obtaining a continuous series of current pressure snapshot values, the next step is to use this historical data to train a pre-deployed time series prediction model to enable it to predict future pressure. The principle of this process is to feed the model with a large number of historical pressure snapshot value time series (for example, every ten minutes in the past few months). The Long Short-Term Memory (LSTM) model is particularly well-suited for this task because its unique gating structure effectively captures long-term dependencies in data sequences, avoiding the vanishing or exploding gradient problems that can plague traditional models when processing long sequences. This training process is essentially an optimization process, where the model continuously adjusts its internal parameters to ensure that its predictions align as closely as possible with real historical data.

[0071] Once the time series forecasting model is trained, it can be used for real-time forecasting. As input, it is fed into the trained model. The model will use its learned internal rules to analyze the current state and calculate a pressure forecast value within a specific time window in the future (for example, the next hour). The output of the future pressure forecast value can be expressed as , where M represents the trained time series prediction model. The core value of this step lies in introducing a forward-looking perspective to stress analysis, moving beyond simply reflecting current congestion to providing an early warning of potential congestion. This foresight enables more proactive and proactive allocation of medical resources and patient guidance, effectively preventing the occurrence or exacerbation of congestion and significantly enhancing the overall system's intelligence and proactive service capabilities.

[0072] Finally, in order to generate a final indicator that is both stable and sensitive to trends, it is necessary to dynamically weight the current pressure snapshot value representing the present and the future pressure forecast value that indicates the future to generate a real-time composite pressure index. This combination is not a simple addition and average, but uses dynamic weights because the importance of the current value and the forecast value is different in different situations. For example, when the forecast value shows that the pressure is about to rise sharply, the weight of the forecast value should be increased to provide an early warning. This dynamic weighted combination process can be achieved through the formula to achieve, where It is a dynamic weight coefficient between 0 and 1, and its value can be adjusted in real time according to factors such as the rate of change of the predicted value and the current time. The final effect of this step is to produce a highly reliable composite index. It not only smoothes out possible instantaneous noise and glitches in current data, but also fully integrates scientific predictions of future trends, making it a more comprehensive, smarter, and more decision-supporting core indicator for measuring the service pressure of medical institutions.

[0073] In one embodiment, mapping the real-time composite pressure index of each medical institution to geographic space to construct a dynamic pressure potential field representing the degree of medical service congestion in the region includes the following steps:

[0074] Each medical institution is abstracted as a virtual mass point in the geographic coordinate system, and the real-time composite pressure index of the medical institution is used as the mass point value of the virtual mass point;

[0075] Based on the geographical locations and particle values ​​of all virtual particles, a spatial interpolation algorithm is used to calculate the pressure potential energy value of each preset grid cell in the geographic coordinate system;

[0076] The pressure potential energy values ​​of all grid cells are aggregated to generate a dynamic pressure potential energy field that covers the entire area and changes continuously.

[0077] In this embodiment, in order to convert discrete hospital pressure data into a regional macro-situation, the first step is to mathematically abstract each medical institution. Specifically, in the two-dimensional coordinate system of the geographic information system, each medical institution is no longer regarded as a complex entity occupying an area, but is simplified to a virtual particle without volume, whose position is the precise geographic coordinates of the institution. The key attribute of this particle is its numerical value, which is set to a previously calculated real-time composite pressure index that can dynamically reflect the service load of the institution. The essence of this step is to convert a complex real-world problem into a standardized spatial data problem, the effect of which is to lay the foundation for subsequent spatial analysis and calculation, so that the pressure levels of different medical institutions can be quantified and compared in the same spatial framework, completing the key conversion from business data to spatial geometric data.

[0078] After all medical institutions are abstracted as mass points with numerical values, it is necessary to infer the pressure conditions at any location in the area based on the information of these known points. This process is achieved through a spatial interpolation algorithm. The basic principle of this algorithm is the first law of geography, that is, things that are close are more closely related. Therefore, the pressure potential energy value of an unknown location can be estimated by taking the weighted average of the pressure values ​​of all known mass points (i.e. medical institutions) around it according to the distance. The closer the medical institution is, the greater its influence (weight). A common implementation method is the inverse distance weighted method (IDW), and its core formula is ,in is the pressure potential energy value of the grid cell to be calculated, is the distance from the cell to the i-th medical institution, and k is a power exponent used to control the decay rate of the distance effect. This step fills the gaps between medical institutions, expanding the discrete point data into continuous surface data, providing algorithmic support for constructing a complete potential energy field.

[0079] Finally, by systematically applying a spatial interpolation algorithm and aggregating all the calculated results, a continuously changing dynamic pressure potential energy field covering the entire target area can be generated. This approach involves first dividing the entire geographic area into a fine grid composed of numerous tiny cells (e.g., squares). Then, for each cell in the grid, the interpolation formula used in the previous step is applied to calculate the pressure potential energy value at its center. Once the potential energy values ​​for all grid cells have been calculated, they are aggregated to form a large data matrix. This matrix can be intuitively rendered as a color-coded heat map, where darker areas (higher potential energy values) represent "high pressure" areas for medical services, while lighter areas (lower potential energy values) represent "low pressure" areas. The resulting visualization creates a dynamic decision-making tool that not only displays current congestion points but also reveals the distribution, spread, and inter-regional impact of pressure, providing a macro, holistic perspective for intelligent traffic diversion and optimal resource allocation.

[0080] In one embodiment, based on the spatiotemporal distribution characteristics of the dynamic pressure potential energy field, a hospital service vector including a dynamic attraction dimension is constructed for other medical institutions in the region, and by calculating the similarity between the patient demand vector and the service vector of each hospital, one or more candidate medical institutions are screened, including the following steps:

[0081] Extract the spatiotemporal distribution characteristics of the dynamic pressure potential energy field. These characteristics include the pressure potential energy value and pressure potential energy gradient at each medical institution. The pressure potential energy gradient characterizes the rate and direction of pressure change at the corresponding location in the dynamic pressure potential energy field and is used to quantify whether the service pressure of a medical institution is easing or increasing.

[0082] Based on the keywords and department information in the service request, a word embedding model is used to generate a patient demand vector;

[0083] For medical institutions in the region, a hospital service vector is constructed based on the service capability characteristics, the inverse of the pressure potential energy value, and the pressure potential energy gradient of the medical institutions, which includes the static service dimension, the dynamic attraction dimension, and the pressure trend dimension.

[0084] The cosine similarity algorithm is used to calculate the similarity score between the patient demand vector and each hospital service vector;

[0085] Arrange in descending order according to the similarity score, and select one or more medical institutions with scores higher than the preset matching threshold as alternative medical institutions.

[0086] In this implementation, to achieve accurate hospital recommendations, the key spatiotemporal distribution characteristics of each medical institution must first be extracted from the dynamic pressure potential energy field. This involves not only directly reading the pressure potential energy value of each institution's geographic location, which represents its current level of congestion, but also calculating the pressure potential energy gradient at that location. The gradient is a vector calculated by analyzing the differences in the potential energy values ​​of the surrounding grids. Its direction points to the direction of the fastest pressure increase, and its magnitude indicates the severity of the pressure change. This gradient can therefore quantify the future trend of service pressure and clearly indicate whether the institution's congestion is easing or continuing to intensify. By extracting these two features—the current potential energy value and the future trend potential energy gradient—a dynamic and insightful pressure profile is constructed for each medical institution, providing the core basis for subsequent intelligent matching that goes beyond static information.

[0087] Next, the user's vague natural language service request, such as what disease to see and which department to go to, needs to be converted into a standardized data structure that the machine can understand and calculate. This process utilizes the word embedding model. The model has been pre-trained on massive medical text data and has learned the semantic relationship between words. When receiving the keywords and department information in the service request, the model will map each word (such as fever, pediatrics) to a high-dimensional digital vector. By combining these word vectors, such as taking the average, a unique patient demand vector will eventually be generated. The position and direction of this vector in mathematical space accurately encodes the user's core request for medical treatment. This step successfully converts the unstructured text request into a structured numerical vector, laying the foundation for subsequent quantitative similarity calculations.

[0088] In addition to building a patient demand vector, it is also necessary to build a comprehensive, multi-dimensional hospital service vector for each medical institution in the region. . This vector is designed to contain information from three key dimensions. The first is the static service dimension, which is used to characterize the inherent service capabilities of the hospital, such as its key specialties, equipment level, etc. The second is the dynamic attractiveness dimension, which is quantified by the inverse of the pressure potential energy value, that is, the lower the pressure, the higher the attractiveness, which intuitively reflects the convenience of seeking medical treatment. The third is the pressure trend dimension, which is constructed using the pressure potential energy gradient calculated in the previous steps to indicate whether the pressure on the hospital is easing or increasing, thereby reflecting its future service potential. By integrating the information from these three dimensions into one vector, a dynamically updated mathematical portrait is created for each hospital that can comprehensively reflect its service status.

[0089] With the vector representing patient needs and the vector representing the service capabilities of each hospital, the cosine similarity algorithm can be used to accurately calculate the degree of match between the two. The core principle of this algorithm is to calculate the cosine value of the angle between two vectors in multidimensional space. The size of this value has nothing to do with the length of the vector itself, only whether the direction they point to is consistent. The calculation formula is The result, Sim, is a score between -1 and 1. The closer the score is to 1, the more aligned the hospital's service vector is with the patient's demand vector, indicating a closer match between the hospital's service profile and the patient's medical needs. By performing this calculation for each candidate hospital in the region, an objective, quantitative matching score is generated for each hospital, simplifying the complex, multi-dimensional matching problem into a clear score ranking problem.

[0090] After calculating the similarity scores between all candidate medical institutions and the patient's needs, the last step is to filter them based on these scores to determine the final recommendation list. This process first sorts all medical institutions in descending order according to their similarity scores, with the highest scores at the top. However, not all institutions with the same score will be recommended, and a preset matching threshold needs to be set. . Only those medical institutions with similarity scores higher than this threshold are considered qualified candidates. The setting of this threshold plays a role in quality control, which can filter out those options that are ranked high but the actual matching degree is not high. Finally, from this list that has passed the threshold test, one or more medical institutions with the highest ranking are selected as the final alternatives. The effect of this step is to ensure that the recommendations pushed to the user are of high quality and highly relevant, thus completing the entire intelligent screening and matching process.

[0091] In one embodiment, generating a diversion guidance strategy based on the pressure characteristics of the target medical institution and the alternative medical institutions in the dynamic pressure potential energy field, and pushing the diversion guidance strategy to the user terminal includes the following steps:

[0092] Extracting pressure potential energy values ​​of target medical institutions and alternative medical institutions from the dynamic pressure potential energy field;

[0093] Calculate the user's estimated waiting time at the target medical institution and alternative medical institutions based on the pressure potential energy value;

[0094] Combining pressure potential energy value and estimated queue time to generate a structured diversion and guidance strategy;

[0095] The diversion guidance strategy is presented in the form of a pop-up window on the interface for confirming the reservation on the user's user terminal.

[0096] In this implementation, in order to generate targeted diversion recommendations, it is first necessary to accurately extract the pressure potential energy values ​​of the target medical institution currently selected by the user and the alternative medical institutions selected by the system at their respective geographical locations from the dynamic pressure potential energy field covering the entire domain. This process is technically a direct data query operation. By obtaining the geographical coordinates of each medical institution, the pressure potential energy value of the corresponding location is directly indexed and read in the pressure potential energy field data grid that has been generated and stored. This step transforms the macro, visual regional congestion situation into precisely quantified values ​​directly relevant to current user decisions. These values ​​form the basis for subsequent calculations and comparisons, providing the raw data input for transforming the abstract concept of stress into specific user perception indicators.

[0097] After obtaining the pressure potential energy value, the next step is to convert it into an indicator that users can most intuitively understand and perceive, namely, the estimated waiting time. This conversion is not a linear correspondence, but is achieved through a function modeled based on historical data. This function needs to be able to reflect the nonlinear growth of waiting time with pressure. That is, when the pressure is low, the waiting time increases slowly, but when the pressure approaches the saturation point, the waiting time will increase sharply. This function can be expressed as ,in is the estimated waiting time. By substituting the pressure potential energy values ​​of the target and alternative medical institutions into this function, the corresponding estimated waiting time can be calculated. The core effect of this step is that it successfully builds a bridge, translating a professional backend data indicator into frontend information that has a decisive impact on user decision-making, greatly enhancing the persuasiveness of the diversion recommendation.

[0098] After calculating the estimated waiting time separately, it is necessary to organically combine this core information with other auxiliary information to generate a clearly structured and contrasting diversion and guidance strategy. This is not just a list of numbers, but also a strategic organization of information. A structured data object will be created, which not only contains the pressure potential energy value and the estimated waiting time calculated based on it, but also integrates comparative information from other dimensions, such as the hospital's expertise matching, geographical distance, and transportation convenience. By displaying this structured information of the target hospital and the alternative hospitals side by side, a clear comparison list is formed. The effect of implementing this step is to integrate the scattered data points into a logical and focused decision support plan. Through the multi-dimensional information presentation, it helps users quickly weigh the pros and cons, and prepares the content for the final interface presentation.

[0099] The last step is to present the carefully designed structured diversion guidance strategy to the user at the most critical decision-making node. This node is selected as the interface where the user is about to complete the appointment and click the confirmation button. Displaying the strategy through an automatically triggered pop-up window (or modal dialog box) can ensure that the information is fully noticed before the user makes the final decision. The pop-up interface will use a highly visual method, such as using different colors or icons to mark the duration and congestion level, to clearly show the huge difference in estimated waiting time between the original option and the recommended option, and provide a clear action button. The effect of this step is to complete the closed loop of the entire guidance process. By intervening in a friendly manner at the right time, the possibility of users accepting diversion suggestions is maximized, thereby successfully transforming the complex background data analysis results into actual results of regional medical resource optimization.

[0100] In one embodiment, the method further comprises the steps of:

[0101] When a user rejects the diversion and guidance strategy, the user's rejection behavior is defined as a rejection event;

[0102] Capturing and associating contextual parameters when a rejection event occurs, the contextual parameters at least include the patient demand vector, the service capability characteristics of the target medical institution and the alternative medical institutions, the difference in pressure potential energy values ​​between the target medical institution and the alternative medical institutions, and the difference in geographical distance;

[0103] Aggregate multiple rejection events and associated contextual parameters to form a decision friction dataset;

[0104] A decision friction map is constructed based on the decision friction dataset, and the decision friction map is used to dynamically adjust the diversion and guidance strategy.

[0105] In this implementation, in order for the guidance strategy to be self-optimizing, it is first necessary to convert the user's rejection behavior into analyzable data. When a user faces a pop-up recommendation and ultimately chooses to ignore the suggestion and stick to the original option, this behavior is defined as a clear rejection event. Technically, this definition process means that the system's front-end interface needs to be able to capture the user's final confirmation operation and determine whether the operation accepts or rejects the diversion. If it is judged to be a rejection, a specific event signal will be triggered. The effect of this step is to transform a silent user choice into a landmark event with clear business meaning that can be recorded and processed by the back-end system. This lays the foundation for the most basic event source for subsequent in-depth analysis of the underlying reasons behind user decisions and the establishment of a feedback loop that can learn from failures.

[0106] After the rejection event is defined and captured, the next key step is to immediately associate and record the complete context parameters when the event occurred. Because the user's rejection is not an isolated behavior, but a comprehensive judgment based on all available information at the time. Therefore, the system must capture and package a series of key data, which at least includes the patient demand vector representing the user's specific needs; the service capability characteristics of the target medical institution and the recommended alternative medical institution; the difference in pressure potential energy value between the two. (in and are the potential energy values ​​of the target and alternative hospitals respectively); and the difference in geographical distance between the two This set of contextual parameters fully reproduces the information users face when making decisions. This step provides a detailed snapshot of the situation for each rejection, making subsequent analysis less a matter of blind guesswork and more a matter of informed data mining.

[0107] Over time, a large number of rejection events and their associated contextual parameters are continuously captured. To conduct pattern analysis, these isolated event records need to be aggregated to form a dedicated decision friction dataset. Technically, this is a data engineering task: the complete contextual parameters of each rejection event are stored as a record in a dedicated database or data warehouse. Each row in this dataset represents a friction between the system's recommendations and the user's preferences, and each column represents the potential factors that may have contributed to this friction, such as pressure difference, distance difference, and departmental compatibility. This step effectively transforms countless individual, momentary decision behaviors into a macro-level, collective intelligence asset amenable to statistical analysis. Based on this aggregated decision friction dataset, a decision friction map can be constructed that reveals the underlying correlations between rejection causes and can be used to dynamically optimize future diversion strategies. First, an unsupervised learning clustering algorithm is used to analyze the contextual parameters in the dataset, clustering similar rejection scenarios and identifying several typical rejection patterns, such as those that are extremely sensitive to distance or those that are highly loyal to a specific hospital brand. Each pattern is defined as a decision friction node in the map. Then, by analyzing the correlation strength between different types of patient demand vectors and these nodes, the correlation edges between the nodes are constructed. When a new user's service request arrives, the system can query the friction node that best matches their needs in the graph, thereby predicting the most likely reason for the user's rejection. Based on this prediction, the system can dynamically adjust the weight parameters when calculating the similarity of alternative hospitals for the user. For example, if the user is identified as distance-sensitive, the negative weight of the distance factor is increased in the similarity calculation. The effect of this step is to enable the diversion strategy to have the ability to learn and evolve, upgrading from a universal rule engine to an intelligent, personalized guidance system that can understand the user's implicit preferences and avoid potential decision-making friction in advance.

[0108] In one embodiment, constructing a decision friction map based on a decision friction dataset includes the following steps:

[0109] An unsupervised learning clustering algorithm is used to perform cluster analysis on the context parameters in the decision friction dataset and identify various rejection patterns.

[0110] Each identified rejection pattern is defined as a decision friction node;

[0111] Node association edges are constructed by analyzing the association strength between patient demand vectors and decision friction nodes;

[0112] All decision friction nodes and node-associated edges are integrated to form a decision friction map that represents the internal reasons why user groups refuse to divert.

[0113] In this embodiment, in order to reveal the underlying reasons behind users' rejection of intelligent diversion recommendations, it is first necessary to perform unsupervised learning cluster analysis on the aggregated decision friction dataset. This process aims to automatically discover natural groupings in the data without pre-setting any labels. Using clustering algorithms such as K-Means or DBSCAN, each rejection event is treated as a multidimensional data point, whose dimensions are composed of contextual parameters such as pressure difference and distance difference. The algorithm will classify similar rejection events into one category based on the distance between these data points in the feature space. For example, all events that are rejected due to excessive distance will naturally cluster together to form a cluster. The effect of this step is to refine and summarize thousands of seemingly chaotic rejection records into a limited number of typical rejection patterns with clear business meanings, completing the first transition from raw data to abstract patterns. After identifying multiple rejection patterns, the next step is to give each pattern a clear identity within the framework of graph construction.

[0114] Specifically, each cluster derived through cluster analysis is defined as an independent decision friction node. This node conceptually represents a specific user decision preference or obstacle. For example, the center point of a cluster might be characterized by a large distance difference and a small pressure difference. The node corresponding to this cluster can be understood as a highly distance-sensitive node. This step is technically a concept mapping process, which converts statistical clustering results into nodes—the basic elements of graph theory. The effect is to establish clear, discrete operational objects for subsequent analysis, solidifying fuzzy patterns into basic units that can be connected and analyzed in the graph, laying the structural foundation for building a knowledge network.

[0115] After having nodes representing rejection patterns, we need to build node association edges to connect users with these patterns to reflect the tendency of different user groups to produce specific rejection behaviors. The core of this step is to analyze the association strength between different types of patient demand vectors and various decision friction nodes in the original data set. For each decision friction node, we can count the frequency of occurrence of various types of patient demand vectors in all rejection events attributed to the node. By calculation, we can get an association strength score, such as using conditional probability To quantify, Represents the demand vector With node The higher the score, the more likely users with this demand characteristic are to exhibit the rejection pattern represented by this node. This step establishes a probabilistic link from user demand to rejection behavior, giving the graph predictive power.

[0116] Finally, by integrating all the created decision friction nodes and the edges associated with them, we can construct a complete decision friction graph that characterizes the internal reasons why the user group refuses to divert. In terms of technical implementation, this is a process of combining the node set and the edge set into a unified graph structure, which can be used , where N is the set of all decision-friction nodes, E is the set of all edges connecting the nodes, and the weight of each edge is the strength of its connection. This graph is not a static image, but a computable, dynamic knowledge base. The ultimate effect is to create an analytical tool that can comprehensively and deeply reveal the user's decision logic. By querying and traversing this graph, we can gain insight into the decision-making preferences behind different medical needs, enabling triage and guidance strategies to fundamentally understand and adapt to users' true thoughts, achieving true intelligence and personalization.

[0117] In one embodiment, dynamically adjusting the diversion and guidance strategy using the decision friction map includes the following steps:

[0118] When screening alternative medical institutions for a user's new service request, the decision friction graph is searched for the target decision friction node that is most strongly associated with the user's patient demand vector;

[0119] Identify the rejection patterns represented by the target decision friction nodes;

[0120] Dynamic adjustment rules are generated based on the rejection pattern, and are applied to the step of screening alternative medical institutions to dynamically adjust the weight parameters when calculating the similarity between the patient demand vector and the service vectors of each hospital.

[0121] In this embodiment, when a new service request is generated, the system will first use its corresponding patient demand vector to query the pre-built decision friction map. The purpose of this query is to find the decision friction node that is most strongly associated with the current user demand. In specific implementation, the preset association strength score between the patient demand vector and each node in the map will be calculated. This score reflects the probability that users with such demands have historically exhibited a specific rejection pattern. By comparing all scores, the node with the highest score can be locked, that is, the target decision friction node. The selection process can be expressed as ,in is the edge weight representing the strength of the association. The effect of this step is to accurately predict the user’s most likely decision bias or concern before the recommendation is generated.

[0122] After locking in the target decision friction node, the next step is to clarify the specific rejection pattern represented by the node. Since each node is constructed by clustering a class of rejection events with similar contextual parameters, each node naturally carries semantic information about the pattern it represents. This process is an information extraction operation, that is, reading the core features of the target node from its attributes. For example, if the target node is characterized by a huge difference in geographical distance and a small difference in pressure potential energy value, then the rejection pattern it represents is determined to be extremely sensitive to distance. The effect of this step is to convert the abstract node query result in the previous step into a specific and understandable business rule, such as users giving priority to distance rather than queue time, which provides a clear guiding direction for subsequent dynamic adjustment of strategies.

[0123] Finally, based on the determined rejection pattern, a dynamic adjustment rule will be generated immediately and applied to the screening process of alternative medical institutions. This rule will directly act on the link of calculating the similarity between the patient demand vector and the hospital service vector by dynamically adjusting the weights of various parameters. For example, if the user's rejection pattern is determined to be distance-sensitive, the negative weight related to the geographical distance in the hospital service vector will be temporarily and significantly increased when calculating the similarity. The adjusted similarity calculation can be regarded as ,in is a dynamic weight vector generated based on the rejection pattern. The ultimate effect of this step is that the recommendation algorithm can proactively avoid potential user rejection points and generate a highly personalized list of alternatives that better aligns with the user's implicit preferences, significantly improving the success rate of the diversion and guidance strategy and user satisfaction.

[0124] The present invention also discloses a regional medical data analysis and processing system based on big data technology, including a memory, a processor, and a computer program stored on the memory and runnable on the processor. When the processor executes the computer program, it implements the regional medical data analysis and processing method based on big data technology as described in any one of the above embodiments.

[0125] Among them, the processor can adopt a central processing unit (CPU). Of course, according to actual usage, other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. can also be adopted. The general-purpose processor can adopt a microprocessor or any conventional processor, etc., and this application does not impose any restrictions on this.

[0126] Among them, the memory can be an internal storage unit of a computer device, such as a hard disk or memory of a computer device, or an external storage device of a computer device, such as a plug-in hard disk, smart memory card (SMC), secure digital card (SD) or flash memory card (FC) equipped on the computer device. In addition, the memory can also be a combination of an internal storage unit and an external storage device of a computer device. The memory is used to store computer programs and other programs and data required by the computer device. The memory can also be used to temporarily store data that has been output or is to be output. This application does not impose any restrictions on this.

[0127] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of protection of the present application is limited to these examples. In line with the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of different aspects of one or more embodiments of the present application as described above, which are not provided in detail for the sake of simplicity.

[0128] The one or more embodiments of this application are intended to encompass all such substitutions, modifications, and variations that fall within the broad scope of this application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of one or more embodiments of this application should be included in the scope of protection of this application.

Claims

1. A regional medical data analysis and processing method based on big data technology, characterized in that: The steps include: Collect multi-dimensional dynamic data of each medical institution in the region, build a composite pressure index model based on the multi-dimensional dynamic data to predict the service pressure trend of each medical institution, and calculate the real-time composite pressure index of each medical institution; Map the real-time composite pressure index of each medical institution to geographic space to construct a dynamic pressure potential field representing the degree of medical service congestion in the region; When receiving a service request from a user for a target medical institution, determining whether the real-time composite pressure index of the target medical institution exceeds a preset congestion threshold; If the congestion threshold is exceeded, the service request is converted into a patient demand vector. Based on the spatiotemporal distribution characteristics of the dynamic pressure potential field, a hospital service vector with a dynamic attractiveness dimension is constructed for other medical institutions in the area. By calculating the similarity between the patient demand vector and the service vectors of each hospital, one or more alternative medical institutions are selected. A diversion guidance strategy is generated based on the pressure characteristics of the target medical institution and the alternative medical institutions in the dynamic pressure potential energy field, and the diversion guidance strategy is pushed to the user terminal.

2. The regional medical data analysis and processing method based on big data technology according to claim 1 is characterized in that: The method of constructing a composite pressure index model for predicting the service pressure trend of each medical institution based on multi-dimensional dynamic data and calculating the real-time composite pressure index of each medical institution includes the following steps: The static resource data, quasi-real-time business data, and real-time queue data in the multi-dimensional dynamic data are fused into the current pressure snapshot value through a weighted fusion function; Use the historical current pressure snapshot value sequence as training data to train the pre-deployed time series prediction model; Utilize the time series prediction model and output the future pressure forecast value based on the current pressure snapshot value; The current pressure snapshot value and the future pressure forecast value are dynamically weighted and combined to generate a real-time composite pressure index.

3. The regional medical data analysis and processing method based on big data technology according to claim 2 is characterized in that: The time series prediction model is a long short-term memory network model.

4. The regional medical data analysis and processing method based on big data technology according to claim 1 is characterized in that: Mapping the real-time composite pressure index of each medical institution to the geographic space to construct a dynamic pressure potential energy field representing the degree of medical service congestion in the region includes the following steps: Each medical institution is abstracted as a virtual mass point in the geographic coordinate system, and the real-time composite pressure index of the medical institution is used as the mass point value of the virtual mass point; Based on the geographical locations and particle values ​​of all virtual particles, a spatial interpolation algorithm is used to calculate the pressure potential energy value of each preset grid cell in the geographic coordinate system; The pressure potential energy values ​​of all grid cells are aggregated to generate a dynamic pressure potential energy field that covers the entire area and changes continuously.

5. The regional medical data analysis and processing method based on big data technology according to claim 1 is characterized in that: The process of constructing a hospital service vector including a dynamic attraction dimension for other medical institutions in the region based on the spatiotemporal distribution characteristics of the dynamic pressure potential energy field, and screening one or more candidate medical institutions by calculating the similarity between the patient demand vector and the service vectors of each hospital, comprises the following steps: Extract the spatiotemporal distribution characteristics of the dynamic pressure potential energy field. These characteristics include the pressure potential energy value and pressure potential energy gradient at each medical institution. The pressure potential energy gradient characterizes the rate and direction of pressure change at the corresponding location in the dynamic pressure potential energy field and is used to quantify whether the service pressure of a medical institution is easing or increasing. Based on the keywords and department information in the service request, a word embedding model is used to generate a patient demand vector; For medical institutions in the region, a hospital service vector is constructed based on the service capability characteristics, the inverse of the pressure potential energy value, and the pressure potential energy gradient of the medical institutions, which includes the static service dimension, the dynamic attraction dimension, and the pressure trend dimension. The cosine similarity algorithm is used to calculate the similarity score between the patient demand vector and each hospital service vector; Arrange in descending order according to the similarity score, and select one or more medical institutions with scores higher than the preset matching threshold as alternative medical institutions.

6. The regional medical data analysis and processing method based on big data technology according to claim 1 is characterized in that: Generating a diversion guidance strategy based on the pressure characteristics of the target medical institution and the alternative medical institutions in the dynamic pressure potential energy field, and pushing the diversion guidance strategy to the user terminal of the user includes the following steps: Extracting pressure potential energy values ​​of target medical institutions and alternative medical institutions from the dynamic pressure potential energy field; Calculate the user's estimated waiting time at the target medical institution and alternative medical institutions based on the pressure potential energy value; Combining pressure potential energy value and estimated queue time to generate a structured diversion and guidance strategy; The diversion guidance strategy is presented in the form of a pop-up window on the interface for confirming the reservation on the user's user terminal.

7. The regional medical data analysis and processing method based on big data technology according to claim 6 is characterized in that: The method further comprises the steps of: When a user rejects the diversion and guidance strategy, the user's rejection behavior is defined as a rejection event; Capturing and associating contextual parameters when a rejection event occurs, the contextual parameters at least include the patient demand vector, the service capability characteristics of the target medical institution and the alternative medical institutions, the difference in pressure potential energy values ​​between the target medical institution and the alternative medical institutions, and the difference in geographical distance; Aggregate multiple rejection events and associated contextual parameters to form a decision friction dataset; A decision friction map is constructed based on the decision friction dataset, and the decision friction map is used to dynamically adjust the diversion and guidance strategy.

8. The regional medical data analysis and processing method based on big data technology according to claim 7 is characterized in that: The construction of a decision friction map based on a decision friction dataset comprises the following steps: An unsupervised learning clustering algorithm is used to perform cluster analysis on the context parameters in the decision friction dataset and identify various rejection patterns. Each identified rejection pattern is defined as a decision friction node; Node association edges are constructed by analyzing the association strength between patient demand vectors and decision friction nodes; All decision friction nodes and node-associated edges are integrated to form a decision friction map that represents the internal reasons why user groups refuse to divert.

9. The regional medical data analysis and processing method based on big data technology according to claim 8, characterized in that: The method of dynamically adjusting the diversion guidance strategy by using the decision friction map includes the following steps: When screening alternative medical institutions for a user's new service request, the decision friction graph is searched for the target decision friction node that is most strongly associated with the user's patient demand vector; Identify the rejection patterns represented by the target decision friction nodes; Dynamic adjustment rules are generated based on the rejection pattern, and are applied to the step of screening alternative medical institutions to dynamically adjust the weight parameters when calculating the similarity between the patient demand vector and the service vectors of each hospital.

10. A regional medical data analysis and processing system based on big data technology, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for analyzing and processing regional medical data based on big data technology as described in any one of claims 1 to 9 is implemented.

Images