Registration recommendation method and system based on multi-dimensional information

Abstract

The invention relates to the field of data processing, in particular to a registration recommendation method and system based on multi-dimensional information, and the method comprises the steps: obtaining registration demand information and user information of a user, collecting number allocation information from hospital public information, and comparing the registration demand information with the number allocation information to screen out available registration information. And weighting the evaluation factors of multiple dimensions of the available registration information based on the registration deviations of different user groups. And the weight sequence is adjusted by introducing the corresponding relation among the user information, the symptom description and the number of doctor seeing times. And the weight sequence is adjusted based on the common degree of symptom description and the number of doctor seeing times, so that the recommendation information can be adjusted according to the common degree of the symptom description and the number of doctor seeing times. And the doctor-seeing recommendation better conforms to the preference of the user, and meanwhile, the patient can be drained, so that the situation of resource shortage of the current large hospital is reduced.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a registration recommendation method and system based on multi-dimensional information. Background Art

[0002] The online appointment registration system is based on computer science, network communication and database technology, and uses the Internet platform to provide patients with convenient appointment registration services. The system usually integrates the hospital's HIS (hospital information system), PACS (image archiving and communication system) and other core systems to realize online processing of registration, payment, report query and other services. Patients can make appointments through various channels (such as mobile APP, WeChat applet, website, etc.), and the system will provide patients with appropriate appointment options based on the hospital's scheduling and number source information.

[0003] The existing online appointment registration system requires users to choose the hospital, department, doctor, time period and other information themselves. However, for many users, they do not actually understand the situation of these hospitals, so they generally prefer to choose expert appointments in large hospitals, which leads to a shortage of medical resources. Summary of the invention

[0004] In view of this, an object of the present invention is to provide a registration recommendation method and system based on multi-dimensional information to solve the problems in the background technology.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions: A registration recommendation method based on multi-dimensional information of the present invention comprises the steps of: Obtaining user information and the user's registration demand information, and obtaining the number of appointments of multiple hospitals, wherein the registration demand information includes department demand information, time demand information, symptom description, and the number of visits for the symptom description, and the number of appointments includes the hospital that issues the appointment, the department that issues the appointment, the doctor that can be booked, and the time period when the doctor can be booked; Determining available registration information from the appointment information of the plurality of hospitals based on the registration demand information; Determine evaluation values ​​of multiple evaluation factors for each available registration information, wherein the evaluation factors include hospital reputation, department reputation, doctor reputation, hospital evaluation information, department evaluation information, doctor evaluation information, average waiting time for department visits, average waiting time for doctor visits, and registration fee; Determine the initial weight sequence of multiple evaluation factors of each available registration information based on the user information and a pre-constructed registration preference database, and modify the initial weight sequence based on the symptom description and the number of visits for the symptom description to obtain a target weight sequence, wherein the registration preference database includes weight sequences of multiple user groups; The evaluation values ​​of the multiple evaluation factors are weighted and evaluated based on the target weight sequence to obtain a comprehensive evaluation score for each available registration information, and registration recommendations are made to the user based on the comprehensive evaluation score.

[0006] In one embodiment of the present application, determining available registration information from the number-granting information of the multiple hospitals based on the registration demand information includes: Comparing the registration demand information with the number-granting information of the multiple hospitals; The appointment information of the multiple hospitals that meets the registration requirement information is used as available registration information.

[0007] In one embodiment of the present application, determining the evaluation values ​​of multiple evaluation factors of each available registration information includes: Obtain search volume for multiple hospitals from search engines and medical information platforms , search volume of departments in multiple hospitals , the number of searches for doctors in multiple hospital departments , evaluation information of multiple hospitals, evaluation information of multiple hospital departments, and evaluation information of doctors in multiple hospital departments; and extracting the average waiting time for multiple doctors before the target time point from the management systems of multiple hospitals ,in, Indicates hospital, Indicates the department of the hospital. It indicates the doctors in the hospital department; Based on the pre-constructed scoring model, multiple evaluation texts in the evaluation information of the multiple hospitals, multiple evaluation texts in the evaluation information of the multiple hospitals' departments, and multiple evaluation texts in the evaluation information of the doctors of the multiple hospitals' departments are evaluated to obtain scores for the multiple evaluation texts; the average score of the multiple evaluation texts is calculated to obtain the evaluation scores of the multiple hospitals. , Evaluation scores of departments in multiple hospitals And the evaluation scores of doctors in multiple hospital departments ; The number of searches for the hospitals , the search volume of the departments of the multiple hospitals , the number of searches for doctors in the departments of the multiple hospitals , the evaluation scores of the multiple hospitals , the evaluation scores of the departments of the multiple hospitals , the evaluation scores of the doctors in the departments of the multiple hospitals , the average waiting time for multiple doctors The registration fees of multiple doctors are normalized respectively to obtain the evaluation values ​​of multiple evaluation factors.

[0008] In one embodiment of the present application, based on a pre-constructed scoring model, multiple evaluation texts in the evaluation information of the multiple hospitals, multiple evaluation texts in the evaluation information of the multiple hospitals' departments, and multiple evaluation texts in the evaluation information of the doctors of the multiple hospitals' departments are evaluated to obtain scores of the multiple evaluation texts, including: Preprocessing the evaluation text to obtain a standardized text, wherein the preprocessing method includes text cleaning and word segmentation; Emotional entities are matched from the standardized text based on a pre-built entity library; and the standardized text containing the emotional entities is input into a pre-trained scoring model to obtain a score for the evaluation text, wherein the evaluation model is obtained by training an artificial neural network with evaluation text samples that have been scored and annotated.

[0009] In one embodiment of the present application, the method for constructing the registration bias database includes: Acquire a plurality of registration sample data after desensitization processing, wherein the registration sample data includes registration user information and registration records; Determine the values ​​of multiple evaluation factor samples and multiple rating factor samples in each registration sample data, and normalize the values ​​of the multiple evaluation factor samples to obtain a registration sample data set; Combining any two evaluation factors and the identity factor in the identity information, and extracting a plurality of two-dimensional data from the registration sample data set based on the combination, wherein the two-dimensional data includes evaluation values ​​of the evaluation factors and values ​​of the identity factors; Constructing a two-dimensional coordinate system based on the evaluation factor and the identity information, and mapping the plurality of two-dimensional data into the two-dimensional coordinate system to obtain a plurality of data points in the two-dimensional coordinate system; Slide along the identity information axis of the two-dimensional coordinate system based on a pre-constructed sliding window, and calculate the variance and average of the evaluation values ​​of the normalized evaluation factors of all data points in the sliding window during each sliding; The segment whose variance is less than the preset variance threshold is taken as the target segment, and the weight of the identity information in each segment is determined based on the variance of the target value segment of the identity information, wherein the weight of the identity information in the target segment is , the weight of the identity information in the non-target segment , is the identity factor, is the mean variance of the target segment; A registration preference database is constructed based on multiple combinations of target segments and multiple combinations of weights of each segment.

[0010] In one embodiment of the present application, the initial weight sequence of multiple evaluation factors of each available registration information is determined based on the user information and a pre-built registration preference database, including: Determining values ​​of a plurality of identity factors from the user information; The weights of the multiple evaluation factors corresponding to each identity factor are determined based on the segment into which the value of each identity factor falls, and an initial weight sequence of the multiple evaluation factors is obtained.

[0011] In one embodiment of the present application, the initial weight sequence is modified based on the symptom description and the number of visits for the symptom description to obtain a target weight sequence, including: Extracting a symptom entity or a combination of symptom entities in the symptom description; Determine the frequency of occurrence of the user's symptom entity or symptom entity combination based on a pre-built statistical database ; When the frequency of occurrence of the symptom entity or the combination of symptom entities is less than the preset first frequency threshold, or the number of visits described by the symptom is greater than the preset number threshold, the weights of the hospital reputation, the department reputation and the doctor reputation are increased to obtain the target weight sequence, and the increase amount is proportional to the frequency of occurrence. negative correlation, or, alternatively, the amount of increase is positively correlated with the number of visits described; When the frequency of occurrence of the symptom entity or the combination of symptom entities is greater than the preset second frequency threshold, or the number of visits described by the symptom is zero, the weights of the hospital reputation, the department reputation and the doctor reputation are reduced to obtain a target weight sequence, and the reduction amount is proportional to the frequency of occurrence. Positive correlation, wherein the second frequency threshold is greater than the first frequency threshold.

[0012] In one embodiment of the present application, the process of constructing the statistical database includes: Obtain user information and symptom description samples of multiple reference users; Extracting symptom description entities from the symptom description samples, and removing duplicates from the symptom entities in the symptom description samples, so that each symptom description sample contains only one symptom entity; Defining multiple value ranges for each identity factor, and combining the value ranges of different identity factors into multiple user groups, and dividing the symptom description samples into data units corresponding to the multiple user groups; Calculate the frequency of occurrence of each symptom entity in each data unit , , where Symptom entity in the data unit The number of The number of samples describing the symptoms in the data unit; A statistical database is constructed based on multiple symptom entities of multiple user groups and the occurrence probabilities of multiple symptom entities.

[0013] In one embodiment of the present application, the evaluation values ​​of the multiple evaluation factors are weighted and evaluated based on the target weight sequence to obtain a comprehensive evaluation score for each available registration information, including: The target weight sequence corresponding to each identity factor and the evaluation values ​​of the multiple evaluation factors are weighted and summed to obtain a comprehensive evaluation score for each available registration information.

[0014] The present application also provides a registration recommendation system based on multi-dimensional information, including: An acquisition module is used to acquire user information and the user's registration demand information, and acquire the number of appointments of multiple hospitals, wherein the registration demand information includes department demand information, time demand information, symptom description, and the number of visits for the symptom description, and the number of appointments includes the number-issuing hospital, the number-issuing department, the doctor who can be booked, and the time period when the doctor can be booked; A screening module, configured to determine available registration information from the number-granting information of the multiple hospitals based on the registration demand information; A preliminary evaluation module, used to determine the evaluation values ​​of multiple evaluation factors for each available registration information, wherein the evaluation factors include hospital reputation, department reputation, doctor reputation, hospital evaluation information, department evaluation information, doctor evaluation information, average waiting time for department visits, average waiting time for doctor visits, and registration fee; a bias determination module, for determining an initial weight sequence of multiple evaluation factors of each available registration information based on the user information and a pre-constructed registration bias database, and modifying the initial weight sequence based on the symptom description and the number of visits for the symptom description to obtain a target weight sequence, wherein the registration bias database includes weight sequences of multiple user groups; The recommendation module is used to perform weighted evaluation on the evaluation values ​​of the multiple evaluation factors based on the target weight sequence, obtain a comprehensive evaluation score for each available registration information, and make registration recommendations to the user based on the comprehensive evaluation score.

[0015] The beneficial effects of the present invention are as follows: a registration recommendation method and system based on multi-dimensional information of the present invention obtains the user's registration demand information and user information, collects the release information from the public information of the hospital, and uses the comparison of the registration demand information and the release information to screen out the available registration information. The present application also uses big data to analyze the registration preferences of different user groups, so as to weight the evaluation factors of multiple dimensions of the available registration information. In addition, the correspondence between user information and symptom descriptions and the number of visits is introduced to adjust the weight sequence. The weight sequence is adjusted based on the commonness of the symptom description and the number of visits, so that for some common symptoms and the first visit, the registration information with lower popularity but other evaluation factors is more in line with the user is recommended. For some users with rare symptoms or multiple visits, registration information with higher popularity is recommended. While the medical recommendation is more in line with the user's preferences, it can also divert patients and reduce the current situation of resource shortage in large hospitals. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention will be further described below in conjunction with the accompanying drawings and embodiments: Figure 1 This is an application scenario diagram of a registration recommendation method based on multi-dimensional information shown in an embodiment of the present application; Figure 2 is a flow chart of a registration recommendation method based on multi-dimensional information shown in an embodiment of the present application; Figure 3 A flowchart for constructing the exercise type-weight sequence relationship table and the physical sign reference database in this application; Figure 4 It is a structural diagram of a registration recommendation system based on multi-dimensional information shown in one embodiment of the present application. DETAILED DESCRIPTION

[0019] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0020] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show the layers related to the present invention rather than being drawn according to the number, shape and size of the layers in actual implementation. In actual implementation, the type, quantity and proportion of each layer may be changed arbitrarily, and the layer layout may also be more complicated.

[0021] In the following description, numerous details are discussed to provide a more thorough explanation of embodiments of the present invention; however, it is apparent to one skilled in the art that embodiments of the present invention may be practiced without these specific details.

[0022] Figure 1 is an application scenario diagram of a registration recommendation method based on multi-dimensional information shown in an embodiment of the present application, such as Figure 1 As shown, in this application, the cloud server 110 is used to perform the recommendation process, and the user can access the cloud server through the terminal 120, upload the registration demand information and identity information through the terminal 120, and the cloud server 110 obtains the currently available registration information in real time by accessing the number release public information platform of multiple hospitals. Based on the above information, the cloud server 110 performs analysis and registration recommendation, and returns the recommendation result to the terminal 120.

[0023] In addition, the information of the relevant persons involved in this application has been obtained with full consent and authorization, and the collection, use and processing of relevant information must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0024] Figure 2 is a flowchart of a registration recommendation method based on multi-dimensional information shown in an embodiment of the present application. Figure 2 As shown, a registration recommendation method based on multi-dimensional information in this embodiment may include the following steps: S210, obtaining user information and the user's registration requirement information, and obtaining the number of appointments of multiple hospitals, wherein the registration requirement information includes department requirement information, time requirement information, symptom description, and the number of visits for the symptom description, and the number of appointments includes the appointment hospital, the appointment department, the doctor who can be booked, and the time period when the doctor can be booked; On the public information website of the hospital, the appointment information generally includes the appointment time, doctor's shift information, department information, registration fee, etc. This application mainly extracts and constructs the tree structure data of hospital information-department information-doctor information-appointment time / registration fee. Figure 3 This is a schematic diagram of the number allocation information of a hospital in an embodiment of the present application. The number allocation information in a tree structure is as follows: Figure 3 shown.

[0025] The user information uploaded by the user mainly includes age and gender, and can also include occupation, medical history, etc. The registration demand information mainly includes the appointment time, appointment department, a brief description of the symptoms, and the number of visits for the current symptoms (the default is 0).

[0026] S220, determining available registration information from the appointment information of the multiple hospitals based on the registration demand information; Among them, the available registration information mainly includes the information of appointment time and appointment department in the demand information, including: S221, comparing the registration demand information with the number-granting information of the multiple hospitals; S222, taking the appointment information of the multiple hospitals that meets the registration requirement information as available registration information.

[0027] Specifically, it mainly compares department information and appointment time, such as Figure 3 As shown, for example, a user needs to make an appointment at the ENT department, and the appointment time is approximately around 10 a.m. Based on this information, the available registration information of multiple hospitals can be filtered out from the tree-structured appointment information.

[0028] S230, determining evaluation values ​​of a plurality of evaluation factors for each available registration information, wherein the evaluation factors include hospital reputation, department reputation, doctor reputation, hospital evaluation information, department evaluation information, doctor evaluation information, average waiting time for department visits, average waiting time for doctor visits, and registration fee; In the present application, multiple dimensions of available registration information are evaluated, so it is necessary to quantify the evaluation factors of multiple dimensions, that is, to determine the evaluation values ​​of multiple evaluation factors of each piece of available registration information.

[0029] This application mainly evaluates each available registration information from four dimensions: popularity, reputation, waiting time, and registration fee. Specifically, it includes hospital popularity, department popularity, doctor popularity, hospital evaluation information, department evaluation information, doctor evaluation information, average waiting time for department visits, average waiting time for doctor visits, and registration fee.

[0030] The process of determining the evaluation value is as follows: S231, obtain the search volume of multiple hospitals from search engines and medical information platforms , search volume of departments in multiple hospitals , the number of searches for doctors in multiple hospital departments , evaluation information of multiple hospitals, evaluation information of multiple hospital departments, and evaluation information of doctors in multiple hospital departments; and extracting the average waiting time for multiple doctors before the target time point from the management systems of multiple hospitals ,in, Indicates hospital, Indicates the department of the hospital. It indicates the doctors in the hospital department; The search volume of hospitals, hospital departments, and doctors in hospital departments reflects their popularity. Departments and doctors may have the same name, so when determining the search volume, search information that includes both hospital and department needs to be included in the search volume of hospital departments. , the search information containing hospital-department-doctor is counted into the search volume of doctors in the hospital department .

[0031] Evaluation information is information that reflects word-of-mouth. The waiting time can be directly extracted from the average waiting time for each department or doctor published on the official website, WeChat public account or other social media platforms.

[0032] S232, based on the pre-constructed scoring model, evaluate the multiple evaluation texts in the evaluation information of the multiple hospitals, the multiple evaluation texts in the evaluation information of the multiple hospital departments, and the multiple evaluation texts in the evaluation information of the doctors of the multiple hospital departments to obtain the scores of the multiple evaluation texts; calculate the average score of the multiple evaluation texts to obtain the evaluation scores of the multiple hospitals , Evaluation scores of departments in multiple hospitals And the evaluation scores of doctors in multiple hospital departments ; In this application, in order to quantify the evaluation information, this application uses a scoring model to score the evaluation text. The construction method of the scoring model in this application includes: Collect enough evaluation texts and annotate them to obtain annotated data; The labeled data is then preprocessed, including text cleaning, word segmentation, and vectorization. The purpose of text cleaning is to remove irrelevant characters, punctuation marks, HTML tags, etc.; word segmentation uses an existing word segmentation algorithm, which will not be described here.

[0033] Since the user's attitude is reflected in the emotional entities in the text, this application extracts the emotional entities from multiple words based on the matching method, and vectorizes the emotional entities to construct training data; Then use a recurrent neural network (RNN) for training. During the training process, the data set needs to be divided into a training set, a validation set, and a test set to evaluate the model performance and prevent overfitting. After the training is completed, the scoring model can be obtained.

[0034] The scoring model constructed using the above process in this application performs the following scoring process: Preprocessing the evaluation text to obtain a standardized text, wherein the preprocessing method includes text cleaning and word segmentation; Emotional entities are matched from the standardized text based on a pre-built entity library; and the standardized text containing the emotional entities is input into a pre-trained scoring model to obtain a score for the evaluation text, wherein the evaluation model is obtained by training an artificial neural network with evaluation text samples that have been scored and annotated.

[0035] S233, the number of searches for the multiple hospitals , the search volume of the departments of the multiple hospitals , the number of searches for doctors in the departments of the multiple hospitals , the evaluation scores of the multiple hospitals , the evaluation scores of the departments of the multiple hospitals , the evaluation scores of the doctors in the departments of the multiple hospitals , the average waiting time for multiple doctors The registration fees of multiple doctors are normalized respectively to obtain the evaluation values ​​of multiple evaluation factors.

[0036] Finally, the above indicators are normalized so that the evaluation values ​​of multiple evaluation factors are between [0,1]. The normalization formula is:

[0037] In the formula, is the normalized hospital search volume, is the minimum search volume among multiple hospitals, The largest search volume among multiple hospitals. is the normalized search volume of the department , is the minimum search volume for departments in multiple hospitals, The maximum search volume for departments in multiple hospitals, is the normalized search volume for doctors, Minimum search volume for doctors in departments across multiple hospitals, The largest number of searches for doctors in multiple hospital departments, is the normalized hospital evaluation score, is the minimum evaluation score among multiple hospitals, is the maximum evaluation score among multiple hospitals, is the normalized evaluation score of the hospital department, is the minimum evaluation score of departments in multiple hospitals. is the maximum evaluation score of departments in multiple hospitals. is the normalized evaluation score of doctors in the hospital departments, is the minimum evaluation score of doctors in multiple hospitals, is the maximum evaluation score of doctors in multiple hospitals, is the normalized registration fee, The registration fees for doctors in multiple hospitals, is the minimum registration fee of doctors in multiple hospitals, It is the highest registration fee for doctors in multiple hospitals. is the normalized average waiting time, is the maximum waiting time for multiple doctors, The minimum waiting time for multiple doctors.

[0038] It is worth noting that the higher the values ​​of other evaluation factors are, the more positive their meanings are. Only the higher the value of the average waiting time is, the more negative its meaning is. Therefore, we take the inverse of it to get the evaluation value of the average waiting time.

[0039] For the above doctor search volume or evaluation information, if the basic number is less than 20, just take the average value.

[0040] S240, determining an initial weight sequence of multiple evaluation factors of each available registration information based on the user information and a pre-constructed registration preference database, and modifying the initial weight sequence based on the symptom description and the number of visits for the symptom description to obtain a target weight sequence, wherein the registration preference database includes weight sequences of multiple user groups; In this application, the registration preference database is constructed by analyzing big data to find the registration preferences of different user groups. Thus, a weight sequence representing the registration preference is provided to the current user. For example, young office workers may prefer registration options with shorter waiting times, while older people may prefer hospitals or departments with higher reputations.

[0041] In this application, the construction process of the registration orientation database is as follows: (1) obtaining a plurality of registration sample data after desensitization processing, wherein the registration sample data includes registration user information and registration records; Since we need to find the relationship between user identity information and registration preferences, we need to collect registration sample data containing user information and registration selection information. The registration sample data needs to be anonymized to remove private information such as name, ID number, and medical insurance number.

[0042] (2) Determine the values ​​of multiple evaluation factor samples and multiple rating factor samples in each registration sample data, and normalize the values ​​of the multiple evaluation factor samples to obtain a registration sample data set; For the values ​​of the same type of evaluation factors, this application performs normalization processing to remove the dimensional differences between different types of data. The normalization algorithm adopts the maximum-minimum normalization algorithm. The specific calculation process can be referred to the previous text and will not be repeated here.

[0043] Please refer to the previous text for the evaluation values ​​of the rating factor samples.

[0044] (3) combining any two evaluation factors and the identity factor in the identity information, and extracting a plurality of two-dimensional data from the registration sample data set based on the combination, wherein the two-dimensional data includes evaluation values ​​of the evaluation factors and values ​​of the identity factors; In this application, any evaluation factor and any identity factor are combined to form a two-dimensional data sequence. By analyzing the distribution of the two-dimensional data sequence, information on whether the current evaluation factor is associated with the current identity factor and how closely the association is obtained is obtained.

[0045] For example, a two-dimensional data sequence is constructed using the identity factor-age and the evaluation factor-average waiting time for medical consultation. An example of the two-dimensional data sequence is: ; (4) constructing a two-dimensional coordinate system based on the evaluation factor and the identity information, and mapping the plurality of two-dimensional data into the two-dimensional coordinate system to obtain a plurality of data points in the two-dimensional coordinate system; In the present application, the constructed two-dimensional data is regarded as two-dimensional coordinates and mapped to a two-dimensional coordinate system xy, so that the two-dimensional data can be mapped to different positions of the coordinate system.

[0046] If the identity factor is a discontinuous and small number of attributes such as gender and occupation, then the coordinate points of gender and occupation can be marked on the x-axis of the two-dimensional coordinate system. When extracting the sliding window information later, the step size and width of the sliding window can be set to 1.

[0047] (5) sliding the pre-constructed sliding window along the identity information axis of the two-dimensional coordinate system, and calculating the variance and average of the evaluation values ​​of the normalized evaluation factors of all data points in the sliding window during each sliding; The variance reflects whether the overall distribution of data points in the window is stable. If it is stable enough, it means that the identity factor segment corresponding to the window is related to the evaluation factor. Taking age and average waiting time as an example, if there is an age group where the waiting time is generally stable, it means that the group in this age group generally has a relatively densely distributed waiting time interval, which is one of the registration preferences of this group.

[0048] (6) taking the segment whose variance is less than the preset variance threshold as the target segment, and determining the weight of the identity information in each segment based on the variance of the target value segment of the identity information, wherein the weight of the identity information in the target segment is , the weight of the identity information in the non-target segment , is the identity factor, is the mean variance of the target segment; The above-mentioned segment is a segment composed of continuous windows, or a segment corresponding to a single window.

[0049] This application uses variance to reflect the stability of the values ​​of the data points in the window. If the variance is less than the preset variance threshold, it means that the values ​​in the window are relatively stable. In addition, if the values ​​in the window are stable but maintained at a low level, it means that the index is related to the group, but the weight is low, and the group does not attach much importance to this evaluation factor. On the contrary, the weight is high, so this application uses the variance average and the evaluation factor average in the segment to construct the weight.

[0050]

[0051] In the formula, is the first scale parameter, is the second scale parameter, is the mean value of the variance in the target segment, is the average value of the evaluation factor in the target segment.

[0052] In the above calculation formula, the variance and mean are used to construct the weight to determine the user's registration preference. If it is a non-target segment, it means that the distribution of the evaluation factor values ​​in this segment is chaotic and irregular. This group may have no obvious preference for this rating factor when registering, so the weight can be 0.

[0053] A registration preference database is constructed based on multiple combinations of target segments and multiple combinations of weights of each segment.

[0054] Finally, by using permutations and combinations to conduct a combined analysis of any identity factor and any evaluation factor, we can obtain a weight sequence of registration preferences corresponding to multiple groups representing different identity factors.

[0055] The above process uses big data to analyze the weight sequence representing the registration preferences of each user group in advance. Therefore, after obtaining the user information of the current user, the corresponding weight sequence can be found through comparison, including: S2401, determining values ​​of multiple identity factors from the user information; S2402, determining the weights of multiple evaluation factors corresponding to each identity factor based on the segment into which the value of each identity factor falls, and obtaining an initial weight sequence of multiple evaluation factors.

[0056] Since preference analysis is performed separately for each identity factor in the previous article, multiple sets of initial weight sequences can be obtained based on multiple identity factors in the user information.

[0057] For example: User A includes identity factors: age 45, gender male, occupation X; The corresponding initial weight sequence may be: Age 45: 0.25 (hospital reputation), 0.02 (department reputation), ..., 0.12 (waiting time); Gender: Male: 0.12 (hospital reputation), 0 (department reputation), ..., 0.32 (waiting time); Occupation X: 0 (hospital reputation), 0 (department reputation), ..., 0.22 (waiting time).

[0058] The above initial weight sequence represents the possible registration preferences corresponding to the values ​​of each identity factor of the current user, but the purpose of this application is also to solve the problem of tight medical resources in large hospitals. In order to solve this problem and provide users with appropriate registration suggestions, this application also uses the symptom descriptions uploaded by users and the number of visits corresponding to the symptom descriptions to modify the initial weight sequence.

[0059] In the application, the main consideration is the commonness of the user's symptoms. If it is a common symptom, the weight of popularity will be reduced, so as to divert the user to some hospitals that are not well-known but are capable of handling common diseases. If it is a rare symptom, the weight corresponding to popularity will be increased.

[0060] In addition, if the user is dissatisfied with the medical results and makes another recommendation for registration, the weight corresponding to the popularity will be increased to help the user recommend authoritative doctors from authoritative hospitals.

[0061] Specifically, the correction method for the initial weight sequence includes: S2411, extracting a symptom entity or a combination of symptom entities in the symptom description; The extraction method of symptom entity is: Segment the symptom description; The matching method is used to extract the target words and obtain the symptom entities.

[0062] S2412, determining the occurrence frequency of the user's symptom entity or symptom entity combination based on a pre-built statistical database ; Whether a symptom is common is evaluated based on the user group, so this application pre-constructs a statistical database of the frequency of occurrence of multiple symptom entities of multiple user groups. In this application, the frequency of occurrence of multiple symptom entities of each user group in the statistical data is used to determine the probability of the symptom entity or symptom entity combination of the current user.

[0063] Among them, the frequency of occurrence of symptom entities or symptom entity combinations The probability of a symptom entity or the product of the probabilities of multiple symptom entities.

[0064] S2413, when the frequency of occurrence of the symptom entity or the combination of symptom entities is less than the preset first frequency threshold, or the number of visits described by the symptom is greater than the preset number threshold, increase the weights of the hospital reputation, the department reputation and the doctor reputation to obtain a target weight sequence, and the increase amount is proportional to the frequency of occurrence. negative correlation, or, alternatively, the amount of increase is positively correlated with the number of visits described; When a symptom entity or a combination of symptom entities is rare, or when a user has visited a doctor for this symptom entity or combination of symptom entities multiple times, the weight corresponding to the popularity is increased based on the frequency and number of visits. Increase The mathematical expression is:

[0065] Alternatively, increase the amount The mathematical expression is:

[0066] In the formula, for are proportional coefficients, The number of visits.

[0067] S2414, when the frequency of occurrence of the symptom entity or the combination of symptom entities is greater than a preset second frequency threshold, or the number of visits described by the symptom is zero, the weights of the hospital reputation, the department reputation, and the doctor reputation are reduced to obtain a target weight sequence, and the reduction amount is proportional to the frequency of occurrence. Positive correlation, wherein the second frequency threshold is greater than the first frequency threshold.

[0068] When the symptom entity or symptom entity combination is relatively common, or when the user visits the doctor for the symptom entity or symptom entity combination for the first time, the weight corresponding to the popularity is reduced based on the frequency and number of visits. Reduction The mathematical expression is:

[0069] S250, performing weighted evaluation on the evaluation values ​​of the plurality of evaluation factors based on the target weight sequence to obtain a comprehensive evaluation score for each available registration information, and performing registration recommendation for the user based on the comprehensive evaluation score.

[0070] Specifically, a weighted sum is performed on the target weight sequence corresponding to each identity factor and the evaluation values ​​of the multiple evaluation factors to obtain a comprehensive evaluation score for each available registration information.

[0071] The comprehensive evaluation score of each available registration information The mathematical expression is: In the formula, For the The first identity factor in the target weight sequence The weight of the evaluation factor, For the The evaluation value of an evaluation factor.

[0072] Finally, multiple available registration information are ranked based on the comprehensive evaluation scores, and the top three available registration information are recommended to the user.

[0073] In one embodiment of the present application, the process of constructing the statistical database includes: (1) Obtain user information and symptom description samples of multiple reference users; (2) extracting symptom description entities from the symptom description samples, and removing duplicate symptom entities from the symptom description samples, so that each symptom description sample contains only one symptom entity; (3) defining multiple value ranges for each identity factor, combining the value ranges of different identity factors into multiple user groups, and dividing the symptom description samples into data units corresponding to the multiple user groups; (4) Calculate the frequency of occurrence of each symptom entity in each data unit , , where Symptom entity in the data unit The number of The number of samples describing the symptoms in the data unit; (5) Construct a statistical database based on multiple symptom entities of multiple user groups and the occurrence probability of multiple symptom entities.

[0074] In this application, some user information and symptom description samples disclosed by users are crawled from relevant websites, and then divided into multiple user groups. The frequency of occurrence of each symptom entity is calculated using multiple symptom entities corresponding to data units of multiple user groups.

[0075] The present invention provides a registration recommendation method based on multi-dimensional information, which obtains the user's registration demand information and user information, collects the release information from the hospital's public information, and uses the comparison of the registration demand information and the release information to screen out the available registration information. The present application also uses big data to analyze the registration preferences of different user groups, so as to weight the evaluation factors of multiple dimensions of the available registration information. In addition, the correspondence between user information and symptom descriptions and the number of visits is introduced to adjust the weight sequence. The weight sequence is adjusted based on the commonness of the symptom description and the number of visits, so that for some common symptoms and the first visit, the registration information with lower popularity but other evaluation factors is more in line with the user is recommended. For some users with rare symptoms or multiple visits, registration information with higher popularity is recommended. While the medical recommendation is more in line with the user's preferences, it can also divert patients and reduce the current situation of resource shortage in large hospitals.

[0076] like Figure 4 As shown, the present application also provides a registration recommendation system based on multi-dimensional information, including: An acquisition module is used to acquire user information and the user's registration demand information, and acquire the number of appointments of multiple hospitals, wherein the registration demand information includes department demand information, time demand information, symptom description, and the number of visits for the symptom description, and the number of appointments includes the number-issuing hospital, the number-issuing department, the doctor who can be booked, and the time period when the doctor can be booked; A screening module, configured to determine available registration information from the number-granting information of the multiple hospitals based on the registration demand information; A preliminary evaluation module, used to determine the evaluation values ​​of multiple evaluation factors for each available registration information, wherein the evaluation factors include hospital reputation, department reputation, doctor reputation, hospital evaluation information, department evaluation information, doctor evaluation information, average waiting time for department visits, average waiting time for doctor visits, and registration fee; a bias determination module, for determining an initial weight sequence of multiple evaluation factors of each available registration information based on the user information and a pre-constructed registration bias database, and modifying the initial weight sequence based on the symptom description and the number of visits for the symptom description to obtain a target weight sequence, wherein the registration bias database includes weight sequences of multiple user groups; The recommendation module is used to perform weighted evaluation on the evaluation values ​​of the multiple evaluation factors based on the target weight sequence, obtain a comprehensive evaluation score for each available registration information, and make registration recommendations to the user based on the comprehensive evaluation score.

[0077] The present invention provides a registration recommendation system based on multi-dimensional information, which obtains the user's registration demand information and user information, collects the release information from the hospital's public information, and uses the comparison of the registration demand information and the release information to screen out the available registration information. The present application also uses big data to analyze the registration preferences of different user groups, so as to weight the evaluation factors of multiple dimensions of the available registration information. In addition, the correspondence between user information and symptom descriptions and the number of visits is introduced to adjust the weight sequence. The weight sequence is adjusted based on the commonness of the symptom description and the number of visits, so that for some common symptoms and the first visit, the registration information with lower popularity but other evaluation factors is more in line with the user is recommended. For some users with rare symptoms or multiple visits, registration information with higher popularity is recommended. While the medical recommendation is more in line with the user's preferences, it can also divert patients and reduce the current situation of resource shortage in large hospitals.

[0078] This embodiment also provides an electronic terminal, including: a processor and a memory; The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the terminal executes any one of the methods in this embodiment.

[0079] The computer-readable storage medium in this embodiment can be understood by ordinary technicians in this field: all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to the computer program. The aforementioned computer program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other media that can store program codes.

[0080] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication with each other. The memory is used to store computer programs, the communication interface is used to communicate, and the processor and the transceiver are used to run computer programs so that the electronic terminal executes each step of the above method.

[0081] In this embodiment, the memory may include a random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0082] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0083] In the above-mentioned embodiments, although the present invention has been described in conjunction with the specific embodiments of the present invention, many replacements, modifications and variations of these embodiments will be apparent to those of ordinary skill in the art based on the foregoing description. The embodiments of the present invention are intended to cover all such replacements, modifications and variations falling within the broad scope of the appended claims.

[0084] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A registration recommendation method based on multi-dimensional information, characterized in that: Includes steps: Obtaining user information and the user's registration demand information, and obtaining the number of appointments of multiple hospitals, wherein the registration demand information includes department demand information, time demand information, symptom description, and the number of visits for the symptom description, and the number of appointments includes the hospital that issues the appointment, the department that issues the appointment, the doctor that can be booked, and the time period when the doctor can be booked; Determining available registration information from the appointment information of the plurality of hospitals based on the registration demand information; Determine evaluation values ​​of multiple evaluation factors for each available registration information, wherein the evaluation factors include hospital reputation, department reputation, doctor reputation, hospital evaluation information, department evaluation information, doctor evaluation information, average waiting time for department visits, average waiting time for doctor visits, and registration fee; Determine the initial weight sequence of multiple evaluation factors of each available registration information based on the user information and a pre-constructed registration preference database, and modify the initial weight sequence based on the symptom description and the number of visits for the symptom description to obtain a target weight sequence, wherein the registration preference database includes weight sequences of multiple user groups; The evaluation values ​​of the multiple evaluation factors are weighted and evaluated based on the target weight sequence to obtain a comprehensive evaluation score for each available registration information, and registration recommendations are made to the user based on the comprehensive evaluation score.

2. The registration recommendation method based on multi-dimensional information according to claim 1, characterized in that: Determining available registration information from the appointment information of the multiple hospitals based on the registration demand information includes: Comparing the registration demand information with the number-granting information of the multiple hospitals; The appointment information of the multiple hospitals that meets the registration requirement information is used as available registration information.

3. The registration recommendation method based on multi-dimensional information according to claim 1, characterized in that: Determine the evaluation values ​​of multiple evaluation factors for each available registration information, including: Obtain search volume for multiple hospitals from search engines and medical information platforms , search volume of departments in multiple hospitals , the number of searches for doctors in multiple hospital departments , evaluation information of multiple hospitals, evaluation information of multiple hospital departments, and evaluation information of doctors in multiple hospital departments; and extracting the average waiting time for multiple doctors before the target time point from the management systems of multiple hospitals ,in, Indicates hospital, Indicates the department of the hospital. It indicates the doctors in the hospital department; Based on the pre-constructed scoring model, multiple evaluation texts in the evaluation information of the multiple hospitals, multiple evaluation texts in the evaluation information of the multiple hospitals' departments, and multiple evaluation texts in the evaluation information of the doctors of the multiple hospitals' departments are evaluated to obtain scores for the multiple evaluation texts; the average score of the multiple evaluation texts is calculated to obtain the evaluation scores of the multiple hospitals. , Evaluation scores of departments in multiple hospitals And the evaluation scores of doctors in multiple hospital departments ; The number of searches for the hospitals , the search volume of the departments of the multiple hospitals , the number of searches for doctors in the departments of the multiple hospitals , the evaluation scores of the multiple hospitals , the evaluation scores of the departments of the multiple hospitals , the evaluation scores of the doctors in the departments of the multiple hospitals , the average waiting time for multiple doctors The registration fees of multiple doctors are normalized respectively to obtain the evaluation values ​​of multiple evaluation factors.

4. The registration recommendation method based on multi-dimensional information according to claim 3, characterized in that: Based on a pre-built scoring model, multiple evaluation texts in the evaluation information of the multiple hospitals, multiple evaluation texts in the evaluation information of the multiple hospitals' departments, and multiple evaluation texts in the evaluation information of the doctors of the multiple hospitals' departments are evaluated to obtain scores of the multiple evaluation texts, including: Preprocessing the evaluation text to obtain a standardized text, wherein the preprocessing method includes text cleaning and word segmentation; Emotional entities are matched from the standardized text based on a pre-built entity library; and the standardized text containing the emotional entities is input into a pre-trained scoring model to obtain a score for the evaluation text, wherein the evaluation model is obtained by training an artificial neural network with evaluation text samples that have been scored and annotated.

5. The registration recommendation method based on multi-dimensional information according to claim 1, characterized in that: The method for constructing the registration preference database includes: Acquire a plurality of registration sample data after desensitization processing, wherein the registration sample data includes registration user information and registration records; Determine the values ​​of multiple evaluation factor samples and multiple rating factor samples in each registration sample data, and normalize the values ​​of the multiple evaluation factor samples to obtain a registration sample data set; Combining any two evaluation factors and the identity factor in the identity information, and extracting a plurality of two-dimensional data from the registration sample data set based on the combination, wherein the two-dimensional data includes evaluation values ​​of the evaluation factors and values ​​of the identity factors; Constructing a two-dimensional coordinate system based on the evaluation factor and the identity information, and mapping the plurality of two-dimensional data into the two-dimensional coordinate system to obtain a plurality of data points in the two-dimensional coordinate system; Slide along the identity information axis of the two-dimensional coordinate system based on a pre-constructed sliding window, and calculate the variance and average of the evaluation values ​​of the normalized evaluation factors of all data points in the sliding window during each sliding; The segment whose variance is less than the preset variance threshold is taken as the target segment, and the weight of the identity information in each segment is determined based on the variance of the target value segment of the identity information, wherein the weight of the identity information in the target segment is , the weight of the identity information in the non-target segment , is the identity factor, is the mean variance of the target segment; A registration preference database is constructed based on multiple combinations of target segments and multiple combinations of weights of each segment.

6. The registration recommendation method based on multi-dimensional information according to claim 5, characterized in that: Determining an initial weight sequence of multiple evaluation factors for each available registration information based on the user information and a pre-built registration preference database, including: Determining values ​​of a plurality of identity factors from the user information; The weights of the multiple evaluation factors corresponding to each identity factor are determined based on the segment into which the value of each identity factor falls, and an initial weight sequence of the multiple evaluation factors is obtained.

7. The registration recommendation method based on multi-dimensional information according to claim 1, characterized in that: The initial weight sequence is modified based on the symptom description and the number of visits for the symptom description to obtain a target weight sequence, including: Extracting a symptom entity or a combination of symptom entities in the symptom description; Determine the frequency of occurrence of the user's symptom entity or symptom entity combination based on a pre-built statistical database ; When the frequency of occurrence of the symptom entity or the combination of symptom entities is less than the preset first frequency threshold, or the number of visits described by the symptom is greater than the preset number threshold, the weights of the hospital reputation, the department reputation and the doctor reputation are increased to obtain the target weight sequence, and the increase amount is proportional to the frequency of occurrence. negative correlation, or, alternatively, the amount of increase is positively correlated with the number of visits described; When the frequency of occurrence of the symptom entity or the combination of symptom entities is greater than the preset second frequency threshold, or the number of visits described by the symptom is zero, the weights of the hospital reputation, the department reputation and the doctor reputation are reduced to obtain a target weight sequence, and the reduction amount is proportional to the frequency of occurrence. Positive correlation, wherein the second frequency threshold is greater than the first frequency threshold.

8. The registration recommendation method based on multi-dimensional information according to claim 7, characterized in that: The construction process of the statistical database includes: Obtain user information and symptom description samples of multiple reference users; Extracting symptom description entities from the symptom description samples, and removing duplicates from the symptom entities in the symptom description samples, so that each symptom description sample contains only one symptom entity; Defining multiple value ranges for each identity factor, and combining the value ranges of different identity factors into multiple user groups, and dividing the symptom description samples into data units corresponding to the multiple user groups; Calculate the frequency of occurrence of each symptom entity in each data unit , , where Symptom entity in the data unit The number of The number of samples describing the symptoms in the data unit; A statistical database is constructed based on multiple symptom entities of multiple user groups and the occurrence probabilities of multiple symptom entities.

9. The registration recommendation method based on multi-dimensional information according to claim 6, characterized in that: The evaluation values ​​of the plurality of evaluation factors are weighted and evaluated based on the target weight sequence to obtain a comprehensive evaluation score for each available registration information, including: The target weight sequence corresponding to each identity factor and the evaluation values ​​of the multiple evaluation factors are weighted and summed to obtain a comprehensive evaluation score for each available registration information.

10. A registration recommendation system based on multi-dimensional information, characterized in that: include: An acquisition module is used to acquire user information and the user's registration demand information, and acquire the number of appointments of multiple hospitals, wherein the registration demand information includes department demand information, time demand information, symptom description, and the number of visits for the symptom description, and the number of appointments includes the number-issuing hospital, the number-issuing department, the doctor who can be booked, and the time period when the doctor can be booked; A screening module, configured to determine available registration information from the number-granting information of the multiple hospitals based on the registration demand information; A preliminary evaluation module, used to determine the evaluation values ​​of multiple evaluation factors for each available registration information, wherein the evaluation factors include hospital reputation, department reputation, doctor reputation, hospital evaluation information, department evaluation information, doctor evaluation information, average waiting time for department visits, average waiting time for doctor visits, and registration fee; a bias determination module, for determining an initial weight sequence of multiple evaluation factors of each available registration information based on the user information and a pre-constructed registration bias database, and modifying the initial weight sequence based on the symptom description and the number of visits for the symptom description to obtain a target weight sequence, wherein the registration bias database includes weight sequences of multiple user groups; The recommendation module is used to perform weighted evaluation on the evaluation values ​​of the multiple evaluation factors based on the target weight sequence, obtain a comprehensive evaluation score for each available registration information, and make registration recommendations to the user based on the comprehensive evaluation score.

Images