Pattern recognition-based medical knowledge recommendation methods, electronic devices, and media

Abstract

This invention relates to the field of data processing technology, specifically to a medical knowledge recommendation method, electronic device, and medium based on pattern recognition. The method includes: clustering sample medical cases to obtain multiple sample clusters; determining the performance value of any sample cluster based on the multiple sample clusters; determining the correlation between unclassified medical cases and any sample cluster based on the performance value; determining the target category to which the unclassified medical cases belong based on the correlation and the any sample cluster; and recommending knowledge based on the target category of the input case query information. This method, by more accurately classifying unclassified medical cases, enables more precise medical knowledge recommendations when receiving input case query information.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a medical knowledge recommendation method based on pattern recognition in the technical field of data processing, an electronic device and a medium. BACKGROUND

[0002] Palliative care refers to a practice that takes patients at the end of life and their families as the center and adopts a multidisciplinary cooperation mode. The main content includes pain and other symptom control, comfortable care, psychological, spiritual and social support, etc. Palliative care is a strong practical specialty with strong scene and rich content. The lack of corresponding cases and practice makes nursing professional students unable to have a real hospice experience and difficult to achieve the course teaching goal. With the continuous promotion of medical education cooperation, schools and hospitals actively cooperate to play the teaching function, fully utilize the clinical case resources, and realize the effective link of theory and practice. In order to facilitate the targeted learning of palliative care knowledge by nursing professional students, the relevant cases of palliative care should be reasonably classified and processed.

[0003] The existing technology classifies the care cases mainly by clustering a large number of care cases, so that the care cases with the same nursing condition are classified into a category. However, the care cases cannot be classified due to unclear nursing conditions or special nursing conditions, and there are similar cases or special cases between the care cases recorded for different nursing conditions. When the students query the corresponding nursing condition care cases, the problem of omission will occur, and the systematic consideration of the students' interest and ability is also lacking, which cannot highlight the advantages of personalized teaching. In addition, there is also a lack of interactivity in case teaching evaluation data collection and student weak knowledge point analysis. Therefore, in the process of selecting recommendation data, the state attribution of the cases cannot be accurately divided, and accurate recommendation cannot be performed. SUMMARY

[0004] In order to solve the technical problem that in the prior art, there are similar cases or special cases between the care cases recorded for different nursing conditions, and in the process of selecting recommendation data, the state attribution of the cases cannot be accurately divided, and accurate recommendation cannot be performed, the purpose of the present application is to provide a medical knowledge recommendation method based on pattern recognition, an electronic device and a medium, and the technical scheme adopted is as follows:

[0005] In the first aspect, the present application embodiment provides a medical knowledge recommendation method based on pattern recognition, which comprises:

[0006] Clustering sample care cases to obtain a plurality of sample clustering clusters;

[0007] Determining the performance value of any sample clustering cluster based on the plurality of sample clustering clusters;

[0008] determine a relevance between the unclassified therapeutic case and the any sample clustering cluster based on the performance value;

[0009] determine a target category to which the unclassified therapeutic case belongs based on the relevance and the any sample clustering cluster;

[0010] recommend knowledge based on the target category.

[0011] In a second aspect, a medical knowledge recommendation device based on pattern recognition is provided, and the device comprises:

[0012] a clustering module configured to cluster sample therapeutic cases to obtain a plurality of sample clustering clusters;

[0013] a first determining module configured to determine a performance value of any sample clustering cluster based on the plurality of sample clustering clusters;

[0014] a second determining module configured to determine a relevance between the unclassified therapeutic case and the any sample clustering cluster based on the performance value;

[0015] a third determining module configured to determine a target category to which the unclassified therapeutic case belongs based on the relevance and the any sample clustering cluster;

[0016] a recommendation module configured to recommend knowledge based on the target category.

[0017] In a third aspect, an electronic device is provided, and the device comprises: a memory configured to store executable program code; and a processor configured to call and run the executable program code from the memory, so that the electronic device executes the method in the first aspect or any possible implementation manner of the first aspect.

[0018] In a fourth aspect, a computer program product is provided, and the computer program product comprises: computer program code configured to cause a computer to execute the method in the first aspect or any possible implementation manner of the first aspect when the computer program code is run on the computer.

[0019] In a fifth aspect, a computer readable storage medium is provided, and the computer readable storage medium stores computer program code configured to cause a computer to execute the method in the first aspect or any possible implementation manner of the first aspect when the computer program code is run on the computer.

[0020] The present application has the following beneficial effects: by clustering sample treatment and nursing cases, a plurality of sample clustering clusters are obtained; then, based on the plurality of sample clustering clusters, a performance value of any sample clustering cluster is determined; in this way, the performance value can intuitively reflect whether the nursing situation targeted by each sample clustering cluster is prone to occur. Then, based on the performance value, the correlation between an unclassified treatment and nursing case and the any sample clustering cluster is determined, and the target category to which the unclassified treatment and nursing case belongs is determined by combining the correlation and the any sample clustering cluster; in this way, by analyzing the correlation between the unclassified treatment and nursing case and the any sample clustering cluster obtained after clustering, the belonging state of the unclassified treatment and nursing case can be judged, so that the target category of the unclassified treatment and nursing case can be accurately analyzed. Finally, through the target category of the unclassified treatment and nursing case, the input case query information can be accurately recommended for medical knowledge. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, and the advantages thereof, a brief introduction will be given to the drawings needed in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained from these drawings without creative labor.

[0022] Figure 1 is a schematic diagram of an implementation process of a medical knowledge recommendation method based on pattern recognition provided by an embodiment of the present application;

[0023] Figure 2 is another schematic diagram of an implementation process of a medical knowledge recommendation method based on pattern recognition provided by an embodiment of the present application;

[0024] Figure 3 is still another schematic diagram of an implementation process of a medical knowledge recommendation method based on pattern recognition provided by an embodiment of the present application;

[0025] Figure 4 is still another schematic diagram of an implementation process of a medical knowledge recommendation method based on pattern recognition provided by an embodiment of the present application;

[0026] Figure 5 is a schematic diagram of a structure of a medical knowledge recommendation device based on pattern recognition provided by an embodiment of the present application. DETAILED DESCRIPTION

[0027] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a pattern recognition-based medical knowledge recommendation method, electronic device, and medium proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0028] In the description of the embodiments of the present invention, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of the present invention, "multiple" means two or more.

[0029] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature.

[0030] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0031] The application environment of the pattern recognition-based medical knowledge recommendation method provided by this invention is described in detail below with reference to the accompanying drawings. Please refer to... Figure 1 The diagram illustrates the implementation flow of a pattern recognition-based medical knowledge recommendation method according to an embodiment of the present invention. This method can be implemented through the following steps:

[0032] Step 101: Cluster the sample treatment cases to obtain multiple sample clusters.

[0033] Here, by extracting keywords from sample treatment cases, and using the similarity scores between the extracted sample keywords and classification keywords, a sample space is created. Then, the data within the sample space are clustered to obtain the sample cluster. In some possible implementations, step 101 can be achieved through steps 111 to 114 (not shown in the diagram):

[0034] Step 111: Extract keywords from the sample treatment cases to obtain sample keywords.

[0035] Here, the sample therapy and care case is stored in the database in the form of text, the sample therapy and care case is extracted, and the sentences in the therapy and care case are processed by the maximum matching segmentation algorithm. The key words in the therapy and care case are extracted by the term frequency-inverse document frequency (TF-IDF) score algorithm, so as to obtain the sample key words. The sample key words can include nouns or verbs in the sample therapy and care case for describing the physical condition or psychological state of the user.

[0036] Step 112, determining the similarity between the sample key words and the classification key words.

[0037] Here, the classification key words can be set key words, such as care, pain, psychology, and spirit. The sample key words are compared and analyzed with the multiple classification key words such as care, pain, psychology, and spirit by the word embedding technology, to obtain the similarity score between the sample key words and the classification key words, that is, the similarity between the sample key words and the classification key words.

[0038] Step 113, creating a sample space based on the similarity and the sample therapy and care case.

[0039] Here, the similarity score of the sample therapy and care case and one classification key word is taken as a coordinate axis data to establish the sample space. If the sample therapy and care case only contains one key word, the similarity score of the key word and the classification key word is taken as the similarity score of the therapy and care case and the classification key word, and is put into the sample space. If the sample therapy and care case contains multiple sample key words, the average of the similarity scores of the multiple sample key words and the classification key words is taken as the similarity score of the sample therapy and care case and the classification key word, and is put into the sample space.

[0040] Step 114, clustering the sample space to obtain the multiple sample clustering clusters.

[0041] Here, the multiple data points in the sample space (wherein one data point represents data after a sample nursing case is bound with a similarity score) are clustered by a density clustering algorithm (Density-Based Spatial Clustering of Applications with Noise, DBSCAN), with a radius value of 70 and a number of points in the neighborhood of 20, to obtain multiple clustering clusters. DBSCAN is a representative density-based clustering algorithm. Unlike division and hierarchical clustering methods, DBSCAN defines a cluster as the maximum set of points that are densely connected, can divide areas with sufficiently high density into clusters, and can find clusters of any shape in a noisy spatial database. Compared with the K-means clustering method, the DBSCAN algorithm does not need to know the number of clusters to be formed in advance, and can find clusters of any shape. At the same time, the DBSCAN algorithm can identify noise points. The DBSCAN algorithm is not sensitive to the order of the samples in the database, that is, the input order of the patterns has little effect on the results. However, for the boundary samples between the cluster classes, their attribution may fluctuate according to which cluster class is detected first. In this way, by binding the similarity with the sample nursing case, creating a sample space, and clustering each data in the sample space, the obtained sample clustering clusters can correspond to the similarity, so as to facilitate subsequent classification of unclassified nursing cases through the sample clustering clusters.

[0042] In step 102, based on the multiple sample clustering clusters, a performance value of any sample clustering cluster is determined.

[0043] Here, the performance value of the sample clustering cluster is calculated for each clustering cluster in the multiple sample clustering clusters. The number of nursing cases included in different nursing conditions is different, and the similarity between the nursing cases included in a single nursing condition is also different. When the number of nursing cases included in a single nursing condition is greater, it proves that the current nursing condition is more likely to occur in daily life, and the nursing cases with similar relationships should be classified into this nursing condition. When the relationship between the nursing cases included in a single nursing condition is closer, it proves that the characteristics of the single nursing condition are more obvious, and the nursing cases with such characteristics should be classified into this nursing condition. For example, the classification of a single nursing condition is mainly for the nursing care of patients with psychological diseases, and the nursing cases related to psychology should be classified into this nursing condition. The above situation analyzes the single nursing condition in the clustered sample space, which is mainly for different clusters in the sample space. The number of data points in the cluster, the distribution relationship between the data points, and the size of the space occupied by a single cluster can be used to evaluate a single cluster, that is, a single nursing condition.

[0044] In some possible implementation manners, the step 102 can be implemented by Figure 2 as shown in the following steps:

[0045] In step 201, the distance between each data point and the adjacent data point in the any sample cluster is determined.

[0046] Here, taking the jth cluster as an example, the distance between the ith data point and the adjacent mth data point in the jth cluster is calculated.

[0047] In step 202, the performance value of the any sample cluster is determined based on the distance and the any sample cluster.

[0048] Here, the performance value of the sample cluster is calculated by combining the distance with the corresponding sample cluster, so as to evaluate the occurrence degree of the single nursing condition represented by the sample cluster through the performance value. In this way, by calculating the performance value of each sample cluster, whether the nursing category indicated by each sample cluster is easy to occur can be reflected through the performance value.

[0049] In some possible implementation manners, the step 202 can be implemented by Figure 3 as shown in the following steps:

[0050] In step 301, the space size of the any sample cluster in the sample space is determined.

[0051] Here, the space size is the space size occupied by the any sample cluster in the sample space.

[0052] In step 302, the ratio between the space size and the number of data points in the any sample cluster is determined.

[0053] In step 303, the ratio and the distance are fused to obtain the performance value of the any sample cluster.

[0054] Here, the performance value of the any sample cluster can be obtained by multiplying the ratio and the distance.

[0055] In the steps 301 to 303, taking the jth cluster as an example, the performance value of the jth cluster is as shown in the following formula (1):

[0056] (1);

[0057] wherein, is the distance between the ith data point and the adjacent mth data point in the jth cluster (i.e., the distance between each data point and the adjacent data point). The sum of distances between all data points in the jth cluster class and its adjacent data points. The smaller the sum of distances, the closer the distance between different data points in the jth cluster class, the closer the relationship between the represented nursing cases, and the greater the possibility that the adjacent data points belong to the jth cluster class.

[0058] The number of data points in the jth cluster class (i.e., the number of data points in the sample clustering cluster). The greater the number, the greater the possibility that the nursing situation represented by the current cluster class occurs in daily life, and the greater the possibility that the adjacent data points belong to the jth cluster class. The size of the space occupied by the jth cluster class in the sample space (i.e., the spatial size of the sample clustering cluster in the sample space). The smaller the value, the closer the distance between different data points in the jth cluster class, the closer the relationship between the represented nursing cases, and the greater the possibility that the adjacent data points belong to the jth cluster class. In this way, by combining the ratio between the spatial size of any sample clustering cluster in the sample space and the number of data points in the sample clustering cluster, and the distance between each data point and its adjacent data point, the performance value of the sample clustering cluster can be accurately calculated.

[0059] Step 103, determining the correlation between the unclassified therapeutic nursing case and the any sample clustering cluster based on the performance value.

[0060] Here, the unclassified therapeutic nursing case is a therapeutic nursing case that has not been classified. The any sample clustering cluster can be understood as a classified nursing situation. The correlation between the unclassified therapeutic nursing case and the any sample clustering cluster is the correlation between the unclassified therapeutic nursing case and the classified nursing situation, which is used to represent the distribution relationship of a single therapeutic nursing case and a classified nursing situation in the sample space.

[0061] In some possible implementations, whether the unclassified therapeutic nursing case should be classified into the classified nursing situation (i.e., the sample clustering cluster) should be determined according to the similarity between the content of the unclassified therapeutic nursing case and the classified nursing situation. In the clustered sample space, the performance of a single data point can be represented by the spatial distribution relationship between the single data point and the cluster class. The smaller the distance between the single data point and the center point of the cluster class range, the higher the similarity between the therapeutic nursing case represented by the single data point and the nursing situation represented by the cluster class. Moreover, the greater the performance value of the cluster class, the greater the possibility that the nursing situation represented by the cluster class itself occurs, the greater the attraction degree of the adjacent data points, and the greater the possibility that the therapeutic nursing case represented by the single data point belongs to the nursing situation represented by the cluster class. Therefore, the single data point and the cluster class can be screened by the spatial distribution relationship between the single data point and the cluster class and the performance value of the cluster class. The step 103 can be implemented by the step shown in the following table: Figure 4

[0062] ​Step 401, in the any sample clustering cluster, the spatial distance between the unclassified treatment and protection case and the center point of the any sample clustering cluster is determined.

[0063] Here, taking the jth clustering cluster as an example, the spatial distance between the data point corresponding to the unclassified treatment and protection case and the center point (i.e. cluster center) of the any sample clustering cluster is calculated. In some possible implementation manners, the coordinates of the data point corresponding to the unclassified treatment and protection case in the space and the coordinates of the cluster center in the space can be calculated through the centroid formula, and then the Euclidean distance between the two coordinate values is calculated, that is, the spatial distance between the unclassified treatment and protection case and the center point of the any sample clustering cluster is obtained.

[0064] Step 402, based on the performance value and the spatial distance, the correlation between the unclassified treatment and protection case and the any sample clustering cluster is determined.

[0065] Here, the correlation between the unclassified treatment and protection case and the any sample clustering cluster is calculated by combining the any sample clustering cluster with the performance value. In this way, by combining the spatial distance between the unclassified treatment and protection case and the center point of the any sample clustering cluster (i.e. the classified nursing case) and the performance value of the sample clustering cluster, the correlation between the unclassified treatment and protection case and the any sample clustering cluster can be accurately calculated.

[0066] In some possible implementation manners, the above step 402 can be implemented through the following steps 421 to 423 (not shown in the figure):

[0067] Step 421, based on the performance value and the spatial distance, the consistency between the unclassified treatment and protection case and the any sample clustering cluster is determined.

[0068] Here, taking the ith unclassified treatment and protection case as an example, and the jth sample clustering cluster as an example, the consistency of the ith unclassified treatment and protection case (i.e. the ith unclassified data point) and the jth classified nursing case is calculated. As shown in formula (2):

[0069] (2);

[0070] Wherein, is the performance value of the jth nursing case (jth clustering cluster), and the greater the value is, the more likely it is that the data point close to the jth nursing case (jth clustering cluster) belongs to the cluster. is the distance (i.e. spatial distance) between the ith data point (i.e. the ith unclassified treatment and protection case) and the center point of the jth clustering cluster in the sample space, and is not 0. The smaller the value, the more similar the data features of the i-th data point are to those of the j-th cluster, and the more likely they should be classified into the same category.

[0071] Step 422: Based on the similarity between the unclassified treatment cases and the classification keywords, and the similarity between any sample cluster and the classification keywords, determine the similarity difference.

[0072] Step 423: Based on the consistency and the similarity difference, determine the correlation between the unclassified treatment cases and any sample cluster.

[0073] Here, the relationship between a single nursing care case and the classified nursing care situation (i.e., the sample cluster) is also related to its similarity score with different classification keywords. When the distance between the data point location of a single nursing care case in the sample space and the center point of the classified nursing care situation cluster is close, but the difference in the similarity score between the two locations and a certain classification keyword in the sample space is too large, it proves that the difference between the single nursing care case and the classified nursing care situation in a certain classification bias is too large, and it is not appropriate to classify them into the same category of nursing care situation. The lower the correlation between the single nursing care case and the classified nursing care situation, the better.

[0074] Therefore, we can obtain the correlation between the i-th unclassified care case and the j-th classified care case (i.e., the j-th cluster). As shown in formula (3):

[0075] (3);

[0076] in, This represents the consistency between the i-th unclassified care case and the j-th classified care situation (i.e., the j-th cluster). The larger the value, the more likely the i-th unclassified care case and the j-th classified care situation should be classified into the same category. These are the similarity scores between the keywords of the i-th unclassified nursing case and the j-th classified nursing situation and the m-th classified keywords, respectively. Let be the maximum value of the similarity score difference (i.e., similarity difference) between the center point location of the i-th unclassified nursing case and the j-th classified nursing situation and the different classification keywords, and The value is not zero. The smaller the value, the smaller the difference in the relationship between the i-th unclassified care case and the j-th classified care situation with different classification keywords, the more similar the classification bias, and the more likely they should be classified into the same category. In this way, by using the similarity difference between unclassified care cases, any sample cluster, and classification keywords, and then combining this similarity difference with consistency, the correlation between unclassified care cases and any sample cluster can be accurately analyzed.

[0077] Step 104: Based on the correlation and any sample cluster, determine the target category to which the unclassified treatment case belongs.

[0078] In some possible implementations, step 104 above can be achieved through the following steps 141 and 142 (not shown in the figure):

[0079] Step 141: Based on the correlation, determine the similarity between the unclassified care cases and the classified care situations.

[0080] Here, when a new nursing case is added to an already categorized nursing situation, it affects characteristics such as the similarity of the cases within that categorized situation. When the change in the nursing situation before and after adding a new case is significant, it indicates that the added case is not very similar to the current categorization, making it less likely to belong to the same category. Conversely, the smaller the change in the nursing situation before and after adding a new case, the greater the similarity between the added case and the current categorization, and the greater the likelihood that they belong to the same category. In the clustered sample space, these characteristics can be reflected in the difference in the performance values ​​of individual data points within a single cluster before and after joining it. The smaller the difference in performance values ​​before and after adding a new data point, the more likely the individual data point should belong to a cluster, and the greater the probability that the nursing case represented by the individual data point belongs to the nursing situation represented by that cluster.

[0081] Therefore, we can conclude that the i-th unclassified care case and the j-th classified care situation (i.e., the j-th cluster) are of similarity. As shown in formula (4):

[0082] (4);

[0083] in, The correlation between the i-th unclassified care case and the j-th classified care case is considered. The larger the value, the more likely the i-th unclassified care case should belong to the j-th classified care case. This represents the performance value of the j-th categorized nursing condition. The performance value of the nursing condition (cluster) classification after adding the i-th unclassified treatment case. The difference in performance values ​​between the pre-classified nursing care cases (clusters) after adding the i-th unclassified case is considered. A smaller value indicates a greater similarity in emphasis between the i-th unclassified case and the j-th classified case, suggesting they should belong to the same category. Normalization is performed using an activation function (Sigmoid function) to obtain... , The range of its value is [0, 1].

[0084] Step 142: Based on the homogeneity, determine the target category of the unclassified care cases.

[0085] Here, unclassified nursing cases include those whose distinctions from existing nursing conditions are unclear after clustering, and cases whose nursing conditions are too unique to warrant a separate category. Forcing such cases into the same category as other nursing conditions would lead to confusion in the keyword-based nursing case searches for students later. Nursing cases that should be classified separately often exhibit significant differences from other nursing conditions. In the sample space, this manifests as a small R-value for each data point compared to all other clusters, making it impossible to classify them into any single category. In some possible implementations, the degree of specificity of unclassified nursing cases is calculated by analyzing similarity; and based on this degree of specificity and a preset threshold, the target category of the unclassified nursing case is determined. The degree of specificity of the i-th unclassified nursing case... As shown in formula (5):

[0086] (5);

[0087] in, This represents the similarity between the i-th unclassified care case and the j-th classified care situation (i.e., the j-th cluster). The smaller the value, the less the current i-th unclassified care case belongs to the j-th classified care situation. This represents the number of clusters in the sample space. Let R be the sum of the R values ​​of the i-th unclassified care case and all classified care cases in the sample space. The smaller the R value, the less the i-th unclassified case belongs to any classified care case; the more special it is, the more it should belong to its own category. Not zero. (Yes / No) Normalization is performed using the Sigmoid function to obtain... Its range is [0, 1]. When When the value is greater than 0.5, the i-th unclassified nursing care case is determined to belong to its own nursing condition category. For the remaining unclassified nursing care cases that do not belong to their own nursing condition category, their similarity with all classified nursing conditions is calculated. The category corresponding to the nursing condition with the highest similarity (i.e., the sample cluster) is taken as the target category of the unclassified nursing care case. In this way, by analyzing the similarity between unclassified nursing care cases and classified nursing conditions, it is possible to determine whether an unclassified nursing care case belongs to the same category as any sample cluster, and thus accurately determine the target category of the unclassified nursing care case.

[0088] Step 105: Based on the target category, perform knowledge recommendation on the input case query information.

[0089] Here, based on the above analysis, all treatment cases can be accurately categorized for nursing students to study and search. The categorized data is then written to the memory via a data bus. The electronic device used in this embodiment includes a memory, a processor, a display, and a data bus for communication. When a student enters keywords (i.e., the entered case search information) into the treatment case search system interface, the computer processor will retrieve the corresponding treatment cases from the relevant category by determining the storage unit address, sending the address to the address bus, issuing read / write signals, and transmitting data. These cases will then be displayed on the computer's screen for the student to select.

[0090] In this embodiment of the invention, performance values ​​are calculated for any given sample cluster using multiple sample clusters. These performance values ​​visually indicate whether the nursing conditions targeted by each sample cluster are likely to occur. Then, based on these performance values, the correlation between unclassified care cases and any given sample cluster is determined. By combining this correlation with any given sample cluster, the target category to which the unclassified care case belongs is determined. Thus, by analyzing the correlation between unclassified care cases and any sample cluster obtained after clustering, the status of the unclassified care case can be judged, thereby accurately identifying the target category of the unclassified care case. Finally, based on the target category of the unclassified care case, more accurate knowledge recommendations can be made for the input case query information.

[0091] This invention provides a medical knowledge recommendation device based on pattern recognition, the device comprising:

[0092] The clustering module is used to cluster sample treatment cases to obtain multiple sample clusters;

[0093] The first determining module is used to determine the performance value of any sample cluster based on the plurality of sample clusters;

[0094] The second determining module is used to determine the correlation between unclassified treatment cases and any sample cluster based on the performance value;

[0095] The third determining module is used to determine the target category to which the unclassified treatment case belongs based on the correlation and any sample cluster;

[0096] The recommendation module is used to recommend knowledge based on the target category of the input case query information.

[0097] In some possible implementations, the clustering module is further configured to extract keywords from the sample treatment cases to obtain sample keywords; determine the similarity between the sample keywords and the classification keywords; create a sample space based on the similarity and the sample treatment cases; and cluster the sample space to obtain the multiple sample clusters.

[0098] In some possible implementations, determining the performance value of any sample cluster based on the plurality of sample clusters includes: for any sample cluster, determining the distance between each data point and its neighboring data points; and determining the performance value of any sample cluster based on the distance and the any sample cluster.

[0099] In some possible implementations, the first determining module is further configured to determine the spatial size of any sample cluster in the sample space; determine the ratio between the spatial size and the number of data points in any sample cluster; and fuse the ratio with the distance to obtain the performance value of any sample cluster.

[0100] In some possible implementations, the second determining module is further configured to determine the spatial distance between the unclassified care case and the centroid of the sample cluster in any sample cluster; and to determine the correlation between the unclassified care case and the sample cluster based on the performance value and the spatial distance.

[0101] In some possible implementations, the second determining module is further configured to: determine the consistency between the unclassified care case and any sample cluster based on the performance value and the spatial distance; determine a similarity difference based on the similarity between the unclassified care case and the classification keyword, and the similarity between any sample cluster and the classification keyword; and determine the correlation between the unclassified care case and any sample cluster based on the consistency and the similarity difference.

[0102] In some possible implementations, the third determining module is further configured to determine the similarity between the unclassified care case and any sample cluster based on the correlation; and to determine the target category of the unclassified care case based on the similarity.

[0103] In some possible implementations, the third determining module is further configured to determine the degree of specialness of the unclassified care cases based on the homogeneity; and to determine the target category of the unclassified care cases based on the degree of specialness and a preset degree of specialness threshold.

[0104] Optionally, the transmission medium may be a wired link (e.g., but not limited to, coaxial cable, fiber optic cable, and digital subscriber line (DSL)) or a wireless link (e.g., but not limited to, wireless Fidelity (WIFI), Bluetooth, and mobile device networks).

[0105] It should be noted that the apparatus provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0106] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. For example, as shown... Figure 5 As shown, the device 500 includes a memory 501, a processor 502, and a computer program 503 stored in the memory 501 and running on the processor 502. When the processor 502 executes the computer program 503, the device can perform any of the aforementioned pattern recognition-based medical knowledge recommendation methods. Furthermore, this embodiment of the invention also protects an apparatus that may include a memory and a processor. The memory stores executable program code, and the processor is used to call and execute the executable program code to perform a pattern recognition-based medical knowledge recommendation method provided by this embodiment of the invention.

[0107] This embodiment can divide the device into functional modules based on the above method example. For example, each module can correspond to a specific function, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware. It should be noted that the module division in this embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods. When dividing each module according to a specific function, the device may also include a signal uploading module, a determination module, and an adjustment module, etc. It should be noted that all relevant content of each step involved in the above method embodiment can be referred to in the functional description of the corresponding functional module, and will not be repeated here.

[0108] It should be understood that the apparatus provided in this embodiment is used to execute the above-described pattern recognition-based medical knowledge recommendation method, and therefore can achieve the same effect as the above-described implementation method. When using integrated units, the apparatus may include a processing module and a storage module. When the apparatus is applied to a device, the processing module can be used to control and manage the device's actions. The storage module can be used to support the device in executing mutual program code, etc. The processing module may be a processor or a controller, which can implement or execute various exemplary logic blocks, modules, and circuits as described in this invention. The processor may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of Digital Signal Processing (DSP) and a microprocessor, etc., and the storage module may be a memory.

[0109] Furthermore, the apparatus provided in the embodiments of the present invention may specifically be a chip, component, or module. The chip may include a connected processor and a memory; wherein, the memory is used to store instructions, and when the processor calls and executes the instructions, the chip can execute the pattern recognition-based medical knowledge recommendation method provided in the above embodiments. This embodiment also provides a computer-readable storage medium storing computer program code. When the computer program code is run on a computer, the computer executes the aforementioned method steps to implement the pattern recognition-based medical knowledge recommendation method provided in the above embodiments.

[0110] This embodiment also provides a computer program product. When the computer program product is run on a computer, it causes the computer to perform the aforementioned related steps to realize the pattern recognition-based medical knowledge recommendation method provided in the above embodiment. The device, computer-readable storage medium, computer program product, or chip provided in this embodiment are all used to execute the corresponding methods provided above. Therefore, the beneficial effects they achieve can be referred to in the beneficial effects of the corresponding methods provided above, and will not be repeated here. Through the description of the above embodiments, those skilled in the art can understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In the embodiments provided by this invention, it should be understood that the disclosed device and method can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of modules or units is merely a logical functional division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0111] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous. The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. The above content is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be determined by the protection scope of the claims.

Claims

1. A medical knowledge recommendation method based on pattern recognition, characterized by, The method comprises: Clustering sample therapy cases to obtain a plurality of sample clustering clusters; Based on the plurality of sample clustering clusters, determining the performance value of any sample clustering cluster; Based on the performance value, determining the correlation between the unclassified therapy case and the any sample clustering cluster; Based on the correlation and the any sample clustering cluster, determining the target category to which the unclassified therapy case belongs; Based on the target category, performing knowledge recommendation on input case query information; The clustering of the sample therapy cases to obtain a plurality of sample clustering clusters comprises: Extracting keywords from the sample therapy cases to obtain sample keywords; Determining the similarity between the sample keywords and the classification keywords; Based on the similarity and the sample therapy cases, creating a sample space; Clustering the sample space to obtain the plurality of sample clustering clusters; The determination of the performance value of any sample clustering cluster based on the plurality of sample clustering clusters comprises: For the any sample clustering cluster, determining the distance between each data point and adjacent data points; Based on the distance and the any sample clustering cluster, determining the performance value of the any sample clustering cluster; The determination of the performance value of the any sample clustering cluster based on the distance and the any sample clustering cluster comprises: Determining the spatial size of the any sample clustering cluster in the sample space; Determining the ratio between the spatial size and the number of data points in the any sample clustering cluster; Fusing the ratio and the distance to obtain the performance value of the any sample clustering cluster; The calculation formula of the performance value comprises: wherein, is the representation value of the jth cluster; is the distance between the ith data point in the jth cluster and its adjacent mth data point; is the sum of distances between all data points in the jth cluster and their adjacent data points; is the number of data points in the jth cluster; is the size of the space occupied by the jth cluster in the sample space; Based on the performance value, determining the correlation between the unclassified therapy case and the any sample clustering cluster comprises: In the any sample clustering cluster, determining the spatial distance between the unclassified therapy case and the center point of the any sample clustering cluster; Based on the performance value and the spatial distance, determining the correlation between the unclassified therapy case and the any sample clustering cluster; The determination of the correlation between the unclassified therapy case and the any sample clustering cluster based on the performance value and the spatial distance comprises: Based on the performance value and the spatial distance, determining the consistency between the unclassified therapy case and the any sample clustering cluster; The calculation formula of the consistency comprises: wherein, is the consistency between the ith unclassified therapy case and the jth cluster; is the distance between the ith unclassified therapy case and the center point of the jth cluster in the sample space. Based on the similarity between the unclassified therapy case and the classification keywords, the similarity between the any sample clustering cluster and the classification keywords, determining a similarity difference value; Based on the consistency and the similarity difference value, determining the correlation between the unclassified therapy case and the any sample clustering cluster; The calculation formula of the correlation comprises: wherein, is a correlation between the ith unclassified therapy case and the jth cluster; is a similarity score between the ith unclassified therapy case and the mth classification keyword; is a similarity score between the jth classified care case keyword and the mth classification keyword; is a maximum value of a difference in similarity score between the center point location of the ith unclassified therapy case and the jth classified care case and different classification keywords; is an absolute value sign.

2. The medical knowledge recommendation method based on pattern recognition according to claim 1, characterized in that, The determination of the target category to which the unclassified therapy case belongs based on the correlation and the any sample clustering cluster comprises: Based on the correlation, determining the same category of the unclassified therapy case and the any sample clustering cluster; The calculation formula of the same category comprises: wherein, is the similarity of the ith unclassified case to the jth cluster; is the performance of the jth cluster after the addition of the ith unclassified case. Based on the same category, determining the target category of the unclassified therapy case.

3. The medical knowledge recommendation method based on pattern recognition according to claim 2, characterized in that, The determination of the target category of the unclassified therapy case based on the same category comprises: determine a special degree of the unclassified therapy and nursing case based on the homogeneity; determine a target category of the unclassified therapy and nursing case based on the special degree and a preset special degree threshold.

4. An electronic device, comprising: The electronic device comprises: a memory for storing executable program codes; a processor for calling and running the executable program codes from the memory, so that the electronic device executes the method as claimed in any one of claims 1 to 3. 5.A computer readable storage medium, wherein at least one computer program is stored in the computer readable storage medium, and the computer program is loaded and executed by a processor to implement the method as claimed in any one of claims 1 to 3.

Images