Intelligent registration management system based on big data
Through big data analysis and intelligent allocation, the problems of resource waste and insufficient medical treatment opportunities in the registration management system have been solved, efficient resource utilization and personalized services have been achieved, and medical treatment efficiency and patient satisfaction have been improved.
Patent Information
- Application Number
- CN202510739264.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-04
- Publication Date
- 2025-09-16
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing registration management system lacks big data analysis capabilities and is unable to achieve intelligent and data-based management, resulting in waste of resources and patients' inability to obtain medical treatment opportunities in a timely manner, as well as a lack of accurate predictions and personalized recommendations.
An intelligent registration management system based on big data is used to allocate doctors' free time, department resources and patient appointment times in real time through K-means cluster analysis, reinforcement learning algorithm and deep neural network model. It also combines collaborative filtering algorithm and time series prediction method to provide personalized registration recommendations.
It achieves efficient use of resources, reduces patient waiting time, improves medical efficiency, provides personalized services, and enhances patient satisfaction.
Smart Images
Figure CN120654853A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of big data and intelligent management technology, and specifically to an intelligent registration management system based on big data. Background Art
[0002] Existing registration management systems primarily rely on traditional information technology, with patients registering offline or by phone, or queuing at hospital check-ins. At some large hospitals, patients often need to make appointments days or even weeks in advance, a process largely managed manually. With the advancement of information technology, some hospitals have introduced online registration systems, allowing patients to schedule appointments online, eliminating the need to queue in person at the hospital. Furthermore, existing systems can provide patients with basic information such as doctor schedules and the department being treated, improving the efficiency of traditional registration models. However, these systems are still in their early stages and suffer from numerous technical deficiencies.
[0003] The existing registration management system faces multiple problems. First, after patients make an appointment to register, they often cannot promptly understand the actual medical situation of the hospital, resulting in a waste of resources or a lack of timely medical treatment opportunities. The registration system's information processing capabilities are weak, and it fails to effectively and intelligently allocate patient needs, doctors' free time, and departmental resources. There is often an over-concentration of registration resources or idle waste. Secondly, the existing system lacks big data analysis capabilities and is unable to extract valuable information from patients' historical registration records, medical habits, and other data to make accurate predictions and personalized recommendations. In short, existing technologies have failed to achieve intelligent, data-based, and efficient management of the registration process, and further innovation and optimization are urgently needed. Summary of the Invention
[0004] In response to the shortcomings of the existing technology, the present invention provides an intelligent registration management system based on big data, which solves the problems of intelligent resource allocation and optimization; and accurate prediction and personalized service.
[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: an intelligent registration management system based on big data, comprising:
[0006] Data collection module, used to collect patient registration information, medical records, doctor scheduling information, department resources and patient historical behavior data in real time;
[0007] A data processing module, based on the data collected by the data collection module, clusters the patient group using the K-means clustering analysis algorithm, extracts the patient's historical registration behavior and medical demand characteristics, and provides decision support for the allocation of registration resources;
[0008] The resource allocation module, based on the analysis results of the data processing module, uses a reinforcement learning algorithm to allocate doctors' free time, department resources, and patient appointment times in real time to ensure the rational allocation and efficient use of registration resources;
[0009] The prediction and recommendation module, based on the patient's historical behavior data and combined with a deep neural network model, accurately predicts the patient's future medical needs and recommends the most appropriate registration time, department, and doctor through collaborative filtering algorithms and time series prediction methods, thereby reducing patients' waiting time and improving medical efficiency;
[0010] The user interface module provides a multi-platform interactive interface for patients to make registration appointments, query registration information and receive personalized recommendations.
[0011] Preferably, the patient registration information includes the patient's basic personal information, appointment registration time and department requirements; the medical record includes medical history records, disease type and treatment plan; the doctor scheduling information includes each doctor's idle time and number of patients received; the department resource situation includes department equipment usage and resource idle time.
[0012] Preferably, the K-means cluster analysis algorithm comprises the following steps:
[0013] S3.1 Based on the patient registration data, a K-means algorithm is used to cluster the patients to obtain the types of patient groups;
[0014] S3.2 extracting the patient's medical needs characteristics based on the type of the patient group, wherein the medical needs characteristics include the preference for medical department and the choice of registration time;
[0015] S3.3 generates an optimized allocation plan for registration resources based on the characteristics of the medical needs, so as to provide appropriate registration services for each type of patient. The model formula of the K-means clustering algorithm is as follows:
[0016]
[0017] Among them, J is the sum of squares of cluster variance, k is the number of clusters, C i is the set of all data points of category i, x j is the data point, μ i is the centroid of the ith class.
[0018] Preferably, the reinforcement learning algorithm comprises the following steps:
[0019] S4.1 Based on the clustering results of the data processing module, a reinforcement learning model is constructed to optimize the registration resource allocation strategy by simulating the effects of different resource allocation strategies;
[0020] The reinforcement learning algorithm described in S4.2 uses the Q-learning algorithm to generate resource allocation decisions for each situation, including dynamic optimization of physician schedule adjustments and patient appointment times;
[0021] S4.3 adjusts the allocation of registration resources based on real-time data feedback to ensure that optimal registration resource allocation can be achieved through real-time allocation during peak periods.
[0022] Among them, the model formula of the Q-learning algorithm is as follows:
[0023]
[0024] Among them, Q(s t ,a t ) is the current state s t and action a t The value of α is the learning rate, r t is the immediate reward, γ is the discount factor, is the next state s t+1 The maximum value under .
[0025] Preferably, the deep neural network model comprises the following steps:
[0026] S5.1 Preprocess the patient's historical behavior data to extract the patient's personalized medical characteristics, including common disease types, frequency of medical visits, and department selection;
[0027] S5.2 Predicting the patient's medical needs based on the patient's historical behavioral data and external factors, including seasonal disease peaks, holiday impacts, and public health events, using a deep neural network model;
[0028] S5.3 uses collaborative filtering algorithms to generate personalized recommendations based on patients' time preferences and department resources, and recommends the most suitable registration time, department, and doctor for patients.
[0029] The training model formula of the deep neural network is:
[0030]
[0031] Where X is the input data, W1, W2 are the weight matrices of the network, b1, b2 are bias terms, and σ(·) is the activation function. Predict the results for the model.
[0032] Preferably, the time series prediction method combines a long short-term memory network to perform time series prediction, comprising the following steps:
[0033] S6.1 Use the long short-term memory network model to establish a medical consultation model based on the patient's medical records to capture the temporal characteristics of the patient's medical consultation time;
[0034] S6.2 predict the patient's future appointment time and department needs based on the consultation model, and provide the patient with accurate registration time recommendations;
[0035] S6.3 combines the patient's personalized needs and the doctor's free time to generate multiple registration recommendation plans for the patient to choose from. The model formula of the long short-term memory network model is as follows:
[0036] f t =σ(W f ·[h t-1 ,x t ]+b f )
[0037] i t =σ(W i ·[h t-1 ,x t ]+b i )
[0038]
[0039] o t =σ(W o ·[h t-1 ,x t ]+b o )
[0040] h t =o t *tanh(C t )
[0041] Among them, f t ,i t , C t ,o t ,h t Represent the forget gate, input gate, candidate memory unit, memory unit, output gate and hidden layer output respectively, W f ,W i ,W C ,W o is the weight matrix, b f ,b i ,b C ,b o is the bias term, x t is the current input.
[0042] Preferably, the user interface module further includes: a patient-side interface, a patient-side interface and a push service interface. The patient-side interface supports patients to make registration and appointment operations on mobile terminals or self-service terminals; the hospital management-side interface is used for hospital managers to view patient appointment status, doctor scheduling and department resource allocation; the push service interface pushes personalized registration recommendation information and reminder notifications.
[0043] Preferably, the personalized recommendation plan includes: registration time recommendation, department selection recommendation, doctor recommendation, medical service and project recommendation and health management advice recommendation.
[0044] The present invention provides an intelligent registration management system based on big data. It has the following beneficial effects:
[0045] This big data-based intelligent registration management system uses K-means clustering analysis and reinforcement learning algorithms to dynamically optimize resource allocation, enabling real-time allocation of physician availability, departmental resources, and patient appointment times. Based on patient demographics, treatment needs, and real-time data feedback, the system not only improves hospital resource utilization but also effectively avoids resource waste and overcrowding during peak hours, enhancing operational efficiency and ensuring timely patient care. The application of reinforcement learning algorithms enables automatic policy adjustments based on evolving data during actual operations, ensuring continuous system optimization.
[0046] Accurate predictions and personalized services: This invention combines deep neural network models with time-series prediction methods to accurately predict medical needs based on patients' historical behavioral data and external factors. Based on the patient's personalized needs, the system can recommend the most appropriate registration time, department, and doctor, significantly reducing patients' waiting time and providing personalized services that meet their individual preferences. The system also provides health management advice and medical service recommendations to help patients achieve health management and early prevention, further enhancing their medical experience and satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] Figure 1 It is a flowchart for implementing the invention. DETAILED DESCRIPTION
[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.
[0049] Example 1
[0050] like Figure 1 As shown, an embodiment of the present invention provides an intelligent registration management system based on big data, including a data collection module for real-time collection of patient registration information, medical records, doctor scheduling information, department resource status, and patient historical behavior data. Patient registration information includes basic personal information of the patient, appointment time, and department requirements; medical records include medical history, disease type, and treatment plan; doctor scheduling information includes each doctor's free time and number of patients; department resource status includes department equipment usage and resource idle time.
[0051] The data processing module, based on the data collected by the data acquisition module, clusters the patient group through the K-means clustering analysis algorithm, extracts the patients' historical registration behavior and medical demand characteristics, and provides decision support for the allocation of registration resources.
[0052] The K-means clustering analysis algorithm includes the following steps:
[0053] S3.1 Based on patient registration data, the K-means algorithm is used to cluster patients to obtain the types of patient groups.
[0054] S3.2 Extract the patient's medical treatment demand characteristics based on the type of patient group. The medical treatment demand characteristics include the preference for the medical department and the choice of registration time.
[0055] S3.3 generates an optimized allocation plan for registration resources based on the characteristics of medical needs, so as to provide appropriate registration services for each type of patient. The model formula of the K-means clustering algorithm is as follows:
[0056]
[0057] Among them, J is the sum of squares of cluster variance, k is the number of clusters, C i is the set of all data points of category i, x j is the data point, μ i is the centroid of the ith class.
[0058] The resource allocation module, based on the analysis results of the data processing module, uses a reinforcement learning algorithm to allocate doctors' free time, department resources and patient appointment times in real time to ensure the rational allocation and efficient use of registration resources.
[0059] The reinforcement learning algorithm consists of the following steps:
[0060] S4.1 Based on the clustering results of the data processing module, a reinforcement learning model is constructed to optimize the registration resource allocation strategy by simulating the effects of different resource allocation strategies.
[0061] The S4.2 reinforcement learning algorithm uses the Q-learning algorithm to generate resource allocation decisions for each situation, including dynamic optimization of doctor schedule adjustments and patient appointment times.
[0062] S4.3 adjusts the allocation of registration resources based on real-time data feedback to ensure that optimal registration resource allocation can be achieved through real-time allocation during peak periods.
[0063] Among them, the model formula of the Q-learning algorithm is as follows:
[0064]
[0065] Among them, Q(s t ,a t ) is the current state s t and action a t The value of α is the learning rate, r t is the immediate reward, γ is the discount factor, is the next state s t+1 The maximum value under .
[0066] The prediction and recommendation module, based on patients' historical behavior data and combined with a deep neural network model, accurately predicts patients' future medical needs, and recommends the most appropriate registration time, department, and doctor through collaborative filtering algorithms and time series prediction methods to reduce patients' waiting time and improve medical efficiency.
[0067] The deep neural network model consists of the following steps:
[0068] S5.1 pre-processes the patient's historical behavior data to extract the patient's personalized medical characteristics, which include common disease types, frequency of medical treatment, and department selection.
[0069] S5.2 is based on a deep neural network model and predicts patients' medical needs based on their historical behavioral data and external factors, including seasonal disease peaks, the impact of holidays and public health events.
[0070] S5.3 uses collaborative filtering algorithms to generate personalized recommendations based on patients' time preferences and department resources, and recommends the most suitable registration time, department, and doctor for patients.
[0071] The training model formula of the deep neural network is:
[0072]
[0073] Where X is the input data, W1, W2 are the weight matrices of the network, b1, b2 are bias terms, and σ(·) is the activation function. Predict the results for the model.
[0074] The time series prediction method combines the long short-term memory network to perform time series prediction, including the following steps:
[0075] S6.1 Use the long short-term memory network model to establish a medical consultation model for patient medical records to capture the temporal characteristics of patient medical consultation time.
[0076] S6.2 predicts patients’ future appointment times and department needs based on the appointment model, and provides patients with accurate registration time recommendations.
[0077] S6.3 combines the patient's personalized needs and the doctor's free time to generate multiple registration recommendation plans for the patient to choose from. The model formula of the long short-term memory network model is as follows:
[0078] f t =σ(W f ·[h t-1 ,x t ]+b f )
[0079] i t =σ(W i ·[h t-1 ,x t ]+b i )
[0080]
[0081] o t =σ(W o ·[h t-1 ,x t ]+b o )
[0082] h t =o t *tanh(C t )
[0083] Among them, f t ,i t , C t ,o t ,h t Represent the forget gate, input gate, candidate memory unit, memory unit, output gate and hidden layer output respectively, W f ,W i ,W C ,W o is the weight matrix, b f ,b i ,b C ,b o is the bias term, x t is the current input.
[0084] The specific implementation is as follows:
[0085] The prediction and recommendation module responds to the following scenarios: seasonal disease peak period.
[0086] Condition description: During the flu season, patients tend to increase their demand for treatment for respiratory diseases, especially in winter when colds, flu, coughs and other diseases are prevalent.
[0087] Data input:
[0088] Patient historical behavior data: including the frequency of visits, departments visited, and common disease types in the same season of the previous year.
[0089] External factors: seasonal disease peak period (flu season).
[0090] Department resource situation: Physician scheduling and resource situation in the respiratory and emergency departments. Especially during the flu season, the waiting times in these departments are usually tight.
[0091] Model training:
[0092] Deep neural network models: By combining historical patient behavior data (such as previous appointment times and disease types) with external factors (such as flu season), the model predicts patient demand for medical services during flu season. Through neural network training, the model learns patient patterns during flu season and provides relatively accurate demand forecasts.
[0093] Time Series Prediction (LSTM): The LSTM model captures changes in patients' appointment times during the flu season and predicts patients' registration times in future flu seasons.
[0094] Recommended plan:
[0095] Recommended registration time: Based on patients' historical behavioral data and flu season forecasts, patients are recommended to choose morning or afternoon registration times during the early peak influenza season (e.g., October-December) to avoid peak periods.
[0096] Department recommendation: Patients are recommended to choose the respiratory department, emergency department or infectious department, and avoid choosing overly crowded departments.
[0097] Doctor Recommendation: Recommend a doctor who has experience in managing influenza cases.
[0098] Data support: The frequency of patients' visits to the hospital increases by about 20%-30% during the flu season, and the departments they visit are mainly respiratory and emergency departments.
[0099] The system uses deep neural networks and LSTM models combined with external seasonal disease data to accurately predict patients' medical needs and department selection.
[0100] The user interface module provides a multi-platform interactive interface for patients to make appointments, query appointment information, and receive personalized recommendations. The user interface module further includes: a patient-side interface, a patient-side interface, and a push service interface. The patient-side interface supports patients making appointments on mobile devices or self-service terminals; a hospital management interface allows hospital managers to view patient appointment status, doctor schedules, and department resource allocation; and a push service interface delivers personalized appointment recommendation information and reminder notifications. Personalized recommendations include: appointment time recommendations, department selection recommendations, doctor recommendations, medical service and project recommendations, and health management advice recommendations.
[0101] Example 2
[0102] The difference between this embodiment and the first embodiment is that the prediction and recommendation module responds to the holiday situation.
[0103] Conditions: During long holidays like Spring Festival or National Day, patients may have different medical needs. Some may choose to have a checkup or address common health issues before the holidays, while others may experience emergencies during the holidays.
[0104] Data input:
[0105] Patient historical behavior data: including past holiday visit records.
[0106] External factors: Due to the influence of holidays, patients have a longer period of free time to seek medical treatment, especially the demand for emergency treatment before or after holidays may increase.
[0107] Department resource situation: availability and doctor scheduling of different departments (such as internal medicine, surgery, and emergency department).
[0108] Model training:
[0109] Deep neural network model: The model will be trained by combining patients' historical holiday medical treatment behavior (such as physical examination needs during the Spring Festival or medical records of sudden emergencies) and holiday factors.
[0110] Time Series Prediction (LSTM): The LSTM model analyzes the registration patterns of patients before and after holidays to predict patients' visit times and department needs during the holidays.
[0111] Recommended plan:
[0112] Recommended registration time: Based on patients' holiday preferences and department load, it is recommended that patients choose a registration period a few days before the holiday to avoid the peak queues on or after the holiday.
[0113] Department recommendation: Patients are recommended to choose internal medicine, surgery or emergency department, as the demand for emergency care is high during holidays.
[0114] Doctor recommendation: Recommend a doctor with experience in emergency treatment during holidays to ensure that patients can receive timely treatment.
[0115] Data support: During the past Spring Festival holiday, patients' demand for physical examinations increased by 15%-20%, and the demand for visits to the emergency department and surgery department increased by 30%-50%.
[0116] The system uses a deep neural network model combined with holiday influencing factors to accurately predict patients' registration needs during holidays.
[0117] Example 3
[0118] This embodiment differs from the first embodiment in that the prediction and recommendation module responds to a public health event.
[0119] Condition Description:
[0120] When a public health incident occurs, patients' medical needs will change dramatically, especially the demand for emergency, infectious and respiratory departments will surge.
[0121] Data input:
[0122] Patient historical behavior data: including previous frequency of visits and disease types.
[0123] External factors: such as public health events can significantly affect patients' medical behavior.
[0124] Department resource situation: scheduling and resource situation of the infectious disease department, respiratory department and emergency department.
[0125] Model training:
[0126] Deep neural network model: Combines data from historical public health events with patients' historical medical records to train the model to predict patients' medical needs.
[0127] Time series prediction: The LSTM model is used to capture the time series characteristics of patients' visits and accurately predict their visit times.
[0128] Recommended plan:
[0129] Registration time recommendation: Patients are recommended to choose non-peak hours to avoid crowds.
[0130] Department recommendation: Patients are recommended to choose the infectious disease department or respiratory department.
[0131] Doctor recommendation: Recommend doctors or experts with experience in epidemic prevention, especially in emergency and infectious disease departments.
[0132] Data support: The registration demand for respiratory and infectious diseases departments has increased by 40%-60%.
[0133] The system effectively predicted patients' medical needs by combining deep neural networks and LSTM models, and successfully adjusted the registration recommendation strategy.
[0134] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.
Claims
1. An intelligent registration management system based on big data, characterized in that: include: Data collection module, used to collect patient registration information, medical records, doctor scheduling information, department resources and patient historical behavior data in real time; A data processing module, based on the data collected by the data collection module, clusters the patient group using the K-means clustering analysis algorithm, extracts the patient's historical registration behavior and medical demand characteristics, and provides decision support for the allocation of registration resources; The resource allocation module uses reinforcement learning algorithms to allocate doctors' free time, department resources, and patient appointment times in real time based on the analysis results of the data processing module; The prediction and recommendation module, based on the patient's historical behavior data and combined with a deep neural network model, accurately predicts the patient's future medical needs and recommends the most appropriate registration time, department, and doctor through collaborative filtering algorithms and time series prediction methods; The user interface module provides a multi-platform interactive interface for patients to make registration appointments, query registration information and receive personalized recommendations.
2. The intelligent registration management system based on big data according to claim 1, characterized in that: The patient registration information includes the patient's basic personal information, appointment registration time and department requirements; the medical record includes medical history, disease type and treatment plan; the doctor scheduling information includes each doctor's idle time and number of patients; the department resource situation includes department equipment usage and resource idle time.
3. The intelligent registration management system based on big data according to claim 1, characterized in that: The K-means cluster analysis algorithm includes the following steps: S3.1 Based on the patient registration data, a K-means algorithm is used to cluster the patients to obtain the types of patient groups; S3.2 extracting the patient's medical needs characteristics based on the type of the patient group, wherein the medical needs characteristics include the preference for medical department and the choice of registration time; S3.3 Generate an optimized allocation plan for registration resources based on the characteristics of the medical needs, where the model formula of the K-means clustering algorithm is as follows: Among them, J is the sum of squares of cluster variance, k is the number of clusters, C i is the set of all data points of category i, x j is the data point, μ i is the centroid of the ith class.
4. The intelligent registration management system based on big data according to claim 1, characterized in that: The reinforcement learning algorithm includes the following steps: S4.1 Based on the clustering results of the data processing module, a reinforcement learning model is constructed to optimize the registration resource allocation strategy by simulating the effects of different resource allocation strategies; The reinforcement learning algorithm described in S4.2 uses the Q-learning algorithm to generate resource allocation decisions for each situation, including dynamic optimization of physician schedule adjustments and patient appointment times; S4.3 Adjust the allocation of registration resources based on real-time data feedback, Among them, the model formula of the Q-learning algorithm is as follows: Among them, Q(s t ,a t ) is the current state s t and action a t The value of α is the learning rate, r t is the immediate reward, γ is the discount factor, is the next state s t+1 The maximum value under .
5. The intelligent registration management system based on big data according to claim 1 is characterized by: The deep neural network model includes the following steps: S5.1 Preprocess the patient's historical behavior data to extract the patient's personalized medical characteristics, including common disease types, frequency of medical visits, and department selection; S5.2 Predicting the patient's medical needs based on the patient's historical behavioral data and external factors, including seasonal disease peaks, holiday impacts, and public health events, using a deep neural network model; S5.3 uses collaborative filtering algorithms to generate personalized recommendations based on the patient's time preferences and department resources, and recommends the most suitable registration time, department, and doctor for the patient. The training model formula of the deep neural network is: Where X is the input data, W1, W2 are the weight matrices of the network, b1, b2 are bias terms, and σ(·) is the activation function. Predict the results for the model.
6. The intelligent registration management system based on big data according to claim 1, characterized in that: The time series prediction method combines a long short-term memory network to perform time series prediction, including the following steps: S6.1 Use the long short-term memory network model to establish a medical consultation model based on the patient's medical records to capture the temporal characteristics of the patient's medical consultation time; S6.2 predict the patient's future appointment time and department needs based on the consultation model, and provide the patient with accurate registration time recommendations; S6.3 combines the patient's personalized needs and the doctor's free time to generate multiple registration recommendation plans for the patient to choose from. The model formula of the long short-term memory network model is as follows: f t =σ(W f ·[h t-1 ,x t ]+b f ) i t =σ(W i ·[h t-1 ,x t ]+b i ) the t =σ(W o ·[h t-1 ,x t ]+b o ) h t =o t *tanh(C t ) Among them, f t ,i t , C t ,o t ,h t Represent the forget gate, input gate, candidate memory unit, memory unit, output gate and hidden layer output respectively, W f ,W i ,W C ,W o is the weight matrix, b f ,b i ,b C ,b o is the bias term, x t is the current input.
7. The intelligent registration management system based on big data according to claim 1, characterized in that: The user interface module further includes: a patient-side interface, a hospital management-side interface and a push service interface. The patient-side interface supports patients to make registration and appointment operations on mobile terminals or self-service terminals; the hospital management-side interface is used for hospital managers to view patient appointment status, doctor scheduling and department resource allocation; the push service interface pushes personalized registration recommendation information and reminder notifications.
8. The intelligent registration management system based on big data according to claim 1, characterized in that: The personalized recommendation plan includes: registration time recommendation, department selection recommendation, doctor recommendation, medical service and project recommendation and health management advice recommendation.