Medical project classification methods, devices, equipment and media

Abstract

The present application relates to the field of information processing technology, and provides a method, device, equipment and medium for classifying medical items, the method comprising: obtaining a medical data set, the medical data set comprising medical data of multiple patients, each medical data comprising at least one medical item to be classified for the patient; determining the similarity values ​​between the medical items to be classified based on the medical data set; taking the medical items to be classified as vertices, constructing weighted edges according to the similarity values ​​between the medical items to be classified, and constructing a first complete graph between the medical items to be classified; based on different similarity thresholds, trimming the first complete graph to obtain multiple second complete graphs; determining the classification labels of the medical items to be classified based on the similarity measures between the multiple second complete graphs and the preset medical item classification graph. The embodiment of the present application can complete the classification labels of unclassified medical items based on the classification results of medical item experts.

Description

Technical Field

[0001] This application relates to the field of information processing technology, and in particular to a method, apparatus, electronic device, and computer-readable storage medium for classifying medical items. Background Technology

[0002] The review of medical insurance payments relies heavily on the analysis of detailed medical records from hundreds of millions of patients. However, the medical items in these records often encompass hundreds of thousands of item codes under broad categories such as examinations, surgeries, chemical drugs, and traditional Chinese medicines. To conduct further analysis on such a vast array of items, it is necessary to first classify them.

[0003] However, the classification of medical procedures is not simply a matter of categorizing and labeling. In medical diagnosis and treatment, a single procedure is often applicable to multiple diseases, and combinations of a single procedure with different procedures often have different therapeutic purposes and effects. Therefore, common classification systems start from the therapeutic purpose, defining procedures that are related to each other as a category.

[0004] Currently, the industry primarily relies on classification systems provided by medical experts. However, expert classification methods have certain limitations: they depend on strong medical backgrounds and prior knowledge, making their establishment and maintenance very costly; expert systems rarely cover all items, and detailed data not classified by experts cannot be analyzed subsequently, resulting in the loss of a significant amount of medical record information. Therefore, providing a medical item classification method that can expand the results of item classification and automatically complete the labels for unclassified items has become an urgent technical problem to be solved. Summary of the Invention

[0005] The main objective of this application is to provide a medical item classification method, apparatus, electronic device, and computer-readable storage medium that can supplement the labels of unclassified medical items based on expert medical item classification results.

[0006] To achieve the above objectives, a first aspect of this application proposes a method for classifying medical items, the method comprising:

[0007] Obtain a medical visit dataset, which includes medical visit data of multiple patients, and each medical visit dataset includes at least one medical item to be classified for the patient.

[0008] Based on the medical visit dataset, determine the similarity value between each pair of the medical items to be classified.

[0009] Using the medical items to be classified as vertices, weighted edges are constructed based on the similarity values ​​between each pair of medical items to be classified, thus constructing a first complete graph between the medical items to be classified.

[0010] Based on different similarity thresholds, the first complete graph is cropped to obtain multiple second complete graphs;

[0011] The classification label of the medical item to be classified is determined based on the similarity measure between the multiple second complete maps and the preset medical item classification map.

[0012] According to some embodiments of the present invention, a method for classifying medical items includes determining the classification label of the medical item to be classified based on a similarity measure between the plurality of second complete maps and a preset medical item classification map, comprising:

[0013] Determine the edit distance between the plurality of second complete maps and the medical item classification map;

[0014] Based on the edit distance, a similarity metric is determined between the plurality of second complete graphs and the medical item classification graph;

[0015] Based on the similarity metric, the classification labels for the medical items to be classified are determined.

[0016] According to some embodiments of the medical item classification method provided by the present invention, the step of determining the classification label of the medical item to be classified based on the similarity measure includes:

[0017] The maximum value in the similarity measure is used as the target similarity measure to determine the complete subgraph in the second complete graph corresponding to the target similarity measure;

[0018] Based on the complete subgraph, the classification labels for the medical items to be classified are determined.

[0019] According to some embodiments of the medical item classification method provided by the present invention, the step of determining the similarity value between any two of the medical items to be classified based on the patient visit dataset includes:

[0020] Based on the aforementioned medical visit dataset, the number of patients who have used the medical items to be classified is counted.

[0021] Based on the number of patients, determine the similarity value between each pair of the medical items to be classified.

[0022] According to some embodiments of the medical item classification method provided by the present invention, the similarity value is determined by the following formula:

[0023]

[0024] Wherein, the D abThe similarity value between the medical items to be classified, a and b, is represented by |N(a)|, which represents the number of patients who have used the medical item to be classified, a, and |N(b)| represents the number of patients who have used the medical item to be classified, b, and |N(a)∩N(b)| represents the number of patients who have used both the medical items to be classified, a and b.

[0025] According to some embodiments of the medical item classification method provided by the present invention, the first complete image is cropped based on different similarity thresholds to obtain multiple second complete images, including:

[0026] Based on different similarity thresholds, weighted edges with weights less than the similarity threshold in the first complete graph are deleted to obtain multiple second complete graphs.

[0027] According to some embodiments of the present invention, in the medical item classification method, after cropping the first complete image based on different similarity thresholds to obtain multiple second complete images, and before determining the classification label of the medical item to be classified based on the similarity measure between the multiple second complete images and a preset medical item classification image, the method further includes:

[0028] Based on the second complete graph, an expert classification dataset is obtained, which includes classified medical items with classification labels;

[0029] Using the categorized medical items as vertices, the connection relationships between the vertices are determined based on the categorization labels of the categorized medical items, and a medical item classification graph is constructed among the categorized medical items.

[0030] To achieve the above objectives, a second aspect of this application provides a medical item classification device, the device comprising:

[0031] The acquisition module is used to acquire a medical visit dataset, which includes medical visit data of multiple patients, and each medical visit dataset includes at least one medical item to be classified for the patient's medical visit.

[0032] The processing module is used to determine the similarity value between each pair of the medical items to be classified based on the medical visit dataset;

[0033] A construction module is used to construct a first complete graph between the medical items to be classified, with the medical items to be classified as vertices and based on the similarity values ​​between each pair of the medical items to be classified.

[0034] The cropping module is used to crop the first complete image based on different similarity thresholds to obtain multiple second complete images;

[0035] The classification module is used to determine the classification label of the medical item to be classified based on the similarity measure between the multiple second complete maps and the preset medical item classification map.

[0036] To achieve the above objectives, a third aspect of this application provides an electronic device, the electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the method described in the first aspect.

[0037] To achieve the above objectives, a fourth aspect of the present application provides a storage medium, which is a computer-readable storage medium for computer-readable storage, wherein the storage medium stores one or more computer programs that can be executed by one or more processors to implement the method described in the first aspect.

[0038] This application proposes a method, apparatus, device, and medium for classifying medical items. The method involves acquiring a medical visit dataset containing multiple patient visit data, each data point including at least one medical item to be classified. Based on the dataset, similarity values ​​are determined between pairs of medical items to be classified. Using each medical item as a vertex, weighted edges are constructed based on the similarity values ​​to build a first complete graph. Then, the first complete graph is pruned based on different similarity thresholds to obtain multiple second complete graphs. Classification labels for the medical items to be classified are determined based on a similarity metric between these second complete graphs and a preset medical item classification graph. In this embodiment, labels for unclassified medical items in the medical visit dataset are completed based on the similarity metric and the preset medical item classification graph. Attached Figure Description

[0039] Figure 1 This is a flowchart illustrating a medical item classification method provided in an embodiment of this application;

[0040] Figure 2 This is a flowchart illustrating a medical item classification method provided in an embodiment of this application;

[0041] Figure 3 This is a flowchart illustrating a medical item classification method provided in an embodiment of this application;

[0042] Figure 4 This is a flowchart illustrating a medical item classification method provided in an embodiment of this application;

[0043] Figure 5 This is a flowchart illustrating a medical item classification method provided in an embodiment of this application;

[0044] Figure 6 This is a schematic diagram of the first complete diagram provided in the embodiments of this application;

[0045] Figure 7 This is a schematic diagram of the second complete diagram provided in the embodiments of this application;

[0046] Figure 8 This is a schematic diagram of the medical item classification diagram provided in the embodiments of this application;

[0047] Figure 9 This is a schematic diagram of a second complete diagram provided in another embodiment of this application;

[0048] Figure 10 This is a schematic diagram of a medical item classification diagram provided in another embodiment of this application;

[0049] Figure 11 This is a schematic diagram of the structure of a medical item classification device provided in an embodiment of this application;

[0050] Figure 12 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0052] It should be noted that, 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 application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0053] The review of medical insurance payments relies heavily on the analysis of medical records from hundreds of millions of patients. However, the medical items in these records often encompass hundreds of thousands of item codes under broad categories such as examinations, surgeries, chemical drugs, and traditional Chinese medicines. To conduct further analysis on such a vast array of items, it is necessary to first categorize them.

[0054] However, the classification of medical procedures is not simply a matter of categorizing and labeling. In medical diagnosis and treatment, a single procedure is often applicable to multiple diseases, and combinations of a single procedure with different procedures often have different therapeutic purposes and effects. Therefore, common classification systems start from the therapeutic purpose, defining procedures that are related to each other as a category.

[0055] Currently, the industry mainly relies on classification systems provided by medical experts. However, expert classification methods have certain limitations: expert classification depends on strong medical background and prior knowledge, making its establishment and maintenance very costly; expert systems are difficult to cover all items, and detailed data not classified by experts cannot be analyzed subsequently, resulting in the loss of a large amount of medical record information; therefore, how to provide a medical item classification method that can learn from expert classification results, expand the results of item classification, and automatically complete the labels of unclassified items has become an urgent technical problem to be solved.

[0056] Based on this, embodiments of this application provide a medical item classification method, apparatus, electronic device, and storage medium, which can supplement the labels of unclassified medical items based on expert medical item classification results.

[0057] This application provides a method, apparatus, electronic device, and storage medium for classifying medical items, which will be described in detail through the following embodiments. First, the method for classifying medical items in this application is described.

[0058] Please see Figure 1 , Figure 1 A flowchart illustrating a medical item classification method provided in an embodiment of this application is shown. Figure 1 As shown, the medical item classification method includes, but is not limited to, steps S110 to S150.

[0059] Step S110: Obtain a medical visit dataset, which includes medical visit data of multiple patients, and each medical visit dataset includes at least one medical item to be classified for the patient.

[0060] For example, a patient visit dataset is formed by acquiring visit data from multiple patients, including at least one medical item to be classified.

[0061] Step S120: Based on the medical visit dataset, determine the similarity value between each pair of the medical items to be classified.

[0062] Understandably, based on the medical records of each patient in the medical records dataset, it is possible to identify all medical items to be classified in the medical records dataset, obtain data information about the medical items to be classified in the medical records dataset, and then determine the similarity value between each pair of medical items to be classified.

[0063] For example, the similarity value between pairs of medical items can be determined by counting the number of patients who have used the same medical item to be classified, or by counting the probability of each medical item to be classified appearing under the same diagnosis.

[0064] It should be noted that the similarity value between two unclassified medical items represents the probability that, from the perspective of therapeutic use, the two unclassified medical items belong to the same related medical item definition. For example, based on the patient visit dataset, the similarity value between two unclassified medical items is determined to be between 0 and 100. The higher the similarity value, the greater the probability that, from the perspective of therapeutic use, the two unclassified medical items belong to the same related medical item definition, that is, the more similar the two unclassified medical items are in terms of therapeutic use.

[0065] See Figure 4 , Figure 4 This application provides a flowchart illustrating a medical item classification method according to an embodiment. Figure 4 As shown, in some embodiments, step S120 of the medical item classification method provided in this application includes:

[0066] Step S410: Based on the medical visit dataset, count the number of patients who have used the medical items to be classified.

[0067] It is understandable that the medical visit dataset includes medical visit data of multiple patients, and each medical visit dataset includes at least one medical item to be classified for the patient. Based on this medical visit dataset, the number of patients corresponding to all medical items to be classified in the medical visit data can be counted.

[0068] For example, the medical visit dataset includes medical items to be classified {a,b,c,d,e}. The number of patients who have used medical items a, b, c, d, and e is counted. Then, the number of patients who have used two medical items to be classified, such as the number of patients who have used medical items a and b, is counted.

[0069] Step S420: Based on the number of patients, determine the similarity value between each pair of the medical items to be classified.

[0070] Understandably, the number of patients can include the number of patients who have used one unclassified medical service, the number of patients who have used two unclassified medical services, such as the number of patients who have used unclassified medical service 'a', or the number of patients who have used both unclassified medical services 'a' and 'b'. Based on the definition of the similarity value between pairs of unclassified medical services—that is, the similarity value characterizes the degree of similarity between two unclassified medical services in terms of their therapeutic purpose—the number of patients is used to determine the similarity value between pairs of unclassified medical services.

[0071] In some embodiments, the similarity value is determined by the following formula:

[0072]

[0073] Wherein, the D ab The similarity value between the medical items to be classified, a and b, is represented by |N(a)|, which represents the number of patients who have used the medical item to be classified, a, and |N(b)| represents the number of patients who have used the medical item to be classified, b, and |N(a)∩N(b)| represents the number of patients who have used both the medical items to be classified, a and b.

[0074] For example, if the number of patients who have used medical item a to be classified, |N(a)|, is 200, the number of patients who have used medical item b to be classified, |N(b)|, is 200, and the number of patients who have used both medical items a and b, |N(a)∩N(b)|, is 50, then according to the above formula, the similarity value D between medical items a and b to be classified can be determined. ab It is 0.25.

[0075] Understandably, in the example above, if the number of patients |N(a)∩N(b)| who have used medical items a and b to be classified is 100, then the similarity value D between medical items a and b to be classified is... ab The similarity value is 0.5. As defined, assuming the number of patients who have used both medical procedure a and medical procedure b remains constant, the greater the number of patients who have used both medical procedures a and b, the higher the similarity value determined by the above formula, indicating that medical procedures a and b are more similar in their therapeutic uses. Furthermore, the total number of patients who have used both medical procedures a and b will not exceed the total number of patients who have used either medical procedure a or medical procedure b. In other words, the similarity value determined by the above formula ranges from 0 to 1.

[0076] It should be noted that the similarity value between pairs of medical items to be classified, determined based on the number of patients, can also be calculated using other formulas, such as:

[0077]

[0078] This application does not limit the scope of the embodiments.

[0079] Step S130: Using the medical items to be classified as vertices, construct weighted edges based on the similarity values ​​between each pair of medical items to be classified, and build a first complete graph between the medical items to be classified.

[0080] For example, refer to Figure 6 , Figure 6 A schematic diagram of the first complete diagram provided in an embodiment of this application is shown, as follows: Figure 6As shown, the first complete graph uses the medical items to be classified {a,b,c,d,e} as vertices, and is classified according to the similarity value {D} between each pair of medical items to be classified. ab D ac D ad D ae D bc D bd D be D cd D ce D de Construct weighted edges to build the first complete graph among the medical items {a,b,c,d,e} to be classified.

[0081] In one specific embodiment, the medical item classification method obtains the similarity values ​​between each pair of medical items to be classified, constructs a similarity matrix among all medical items to be classified, and constructs a first complete graph of the medical items to be classified based on the similarity matrix. Here, the elements in the similarity matrix represent the similarity values ​​between each pair of medical items to be classified.

[0082] For example, if the medical items to be classified are {a, b, c, d, e}, then a similarity matrix A is constructed between all the medical items to be classified based on their similarity values.

[0083]

[0084] In this matrix, the elements represent the similarity values ​​between any two medical items to be classified, such as D. cd This represents the similarity value between medical items c and d to be classified.

[0085] It should be noted that in the above embodiments, by constructing a similarity matrix among all medical items to be classified, and using this to map out the first complete graph among all medical items to be classified, it can be understood that the similarity matrix is ​​a data structure used to store the first complete graph, and can also be implemented by other data structures. This application embodiment does not impose any limitations.

[0086] Step S140: Based on different similarity thresholds, the first complete image is cropped to obtain multiple second complete images.

[0087] It is understandable that by defining different similarity thresholds, the first complete graph can be cropped to obtain multiple second complete graphs based on different similarity thresholds.

[0088] In some embodiments, step S140 of the medical item classification method provided in this application includes:

[0089] Based on different similarity thresholds, weighted edges with weights less than the similarity threshold in the first complete graph are deleted to obtain multiple second complete graphs.

[0090] For example, a similarity threshold D is defined. * Furthermore, a higher similarity value indicates that the two medical items to be classified are similar in terms of their therapeutic uses. Therefore, a similarity value D... ab >D * When medical items a and b to be classified are determined to be highly similar, they are classified according to the similarity threshold D. * The first complete graph is cropped, retaining only those with weights greater than the similarity threshold D. * Weighted edges, that is, edges with weights less than the similarity threshold D in the first complete graph, are deleted. * Weighted edges.

[0091] Reference Figure 7 , Figure 7 This illustration shows a schematic diagram of the second complete diagram provided in an embodiment of this application, by deleting, as shown in the diagram. Figure 6 The weighted edges {D} in the first complete graph shown have weights less than the similarity threshold. ac D ad D bd}, to obtain Figure 7 The second complete graph is shown.

[0092] In one specific embodiment, when the first complete graph is obtained through mapping using a similarity matrix (i.e., the first complete graph is stored using a similarity matrix), weighted edges with weights less than the similarity threshold in the first complete graph are deleted based on different similarity thresholds, resulting in multiple second complete graphs, including:

[0093] Based on different similarity thresholds, elements in the similarity matrix that are less than the similarity threshold are deleted, resulting in multiple similarity matrices.

[0094] Understandably, based on different similarity threshold values, weighted edges with weights less than the similarity threshold in the first complete graph are deleted, resulting in multiple second complete graphs.

[0095] For example, the medical items to be classified are {a, b, c, d, e}, and the similarity value between any two medical items to be classified is {D}. ab D ac D ad D ae D bc D bd D be D cd D ce D de Construct the first complete graph G based on a similarity threshold. Delete the first complete graph G whose weights are less than the similarity threshold. Weighted edge {D ad D ce}, thus obtaining the second complete graph G′1; based on the similarity threshold Delete the first complete graph G whose weights are less than the similarity threshold. Weighted edge {D ae D bc}, thus obtaining the second complete graph G′2.

[0096] In one specific embodiment, deleting weighted edges in the first complete graph whose weights are less than the similarity threshold based on different similarity thresholds includes:

[0097] If an isolated vertex exists after deleting a weighted edge, then delete that isolated vertex.

[0098] Understandably, as can be seen from the definition of the similarity threshold, if a medical item to be classified is highly dissimilar to other medical items to be classified, after deleting the weighted edges with weights less than the similarity threshold in the first complete graph, the vertex corresponding to the medical item to be classified becomes an isolated vertex.

[0099] For example, delete such as Figure 6 The weights in the first complete graph shown are less than the similarity threshold D. * Weighted edge {D ab D ac D ad D ae D bd After that, the medical item to be classified, 'a', which is vertex 'a', becomes an isolated vertex. Deleting vertex 'a' yields the following result: Figure 9 The second complete graph shown includes medical items to be classified as {b,c,d,e}.

[0100] Step S150: Determine the classification label of the medical item to be classified based on the similarity measure between the plurality of second complete maps and the preset medical item classification map.

[0101] Understandably, based on the similarity measurement between multiple second complete graphs and the preset medical item classification graph, the second complete graph that is most similar to the preset medical item classification graph is determined as the final classification result of the medical item to be classified, thereby obtaining the classification label of the medical item to be classified.

[0102] It should be noted that the preset medical item classification map is constructed based on the expert classification results of medical items.

[0103] See Figure 5 , Figure 5This application provides a flowchart illustrating a medical item classification method according to an embodiment. Figure 5 As shown, in some embodiments, after step S140 and before step S150, the medical item classification method provided in this application further includes:

[0104] Step S510: Based on the second complete graph, obtain an expert classification dataset, which includes classified medical items with classification labels.

[0105] It should be noted that, based on the second complete graph, that is, based on the medical items to be classified in the second complete graph, an expert classification dataset including classified medical items with classification labels is obtained.

[0106] For example, such as Figure 7 As shown, the unclassified medical items in the second complete graph are {a,b,c,d,e}, and the classified medical items in the obtained expert classification dataset are {a,b,c,d,e}, which correspond one-to-one with the unclassified medical items.

[0107] For example, such as Figure 7 As shown, the unclassified medical items in the second complete graph are {a,b,c,d,e}. The classified medical items {b,c,d,e} in the expert classification dataset are obtained. The classified medical items {b,c,d,e} are a subset of the unclassified medical items {a,b,c,d,e}.

[0108] Step S520: Using the classified medical items as vertices, determine the connection relationships between the vertices based on the classification labels of the classified medical items, and construct a medical item classification graph among the classified medical items.

[0109] For example, such as Figure 7 As shown, the medical items to be classified in the second complete graph are {a,b,c,d,e}, and the classified medical items in the expert classification dataset are {a,b,c,d,e}. Using the classified medical items {a,b,c,d,e} as vertices, the connection relationships between vertices {a,b,c,d,e} are determined based on the classification labels of the classified medical items, constructing the following... Figure 8 The diagram showing the classification of medical items among the categorized medical items is shown.

[0110] For example, such as Figure 7As shown, the medical items to be classified in the second complete graph are {a,b,c,d,e}. Classified medical items {b,c,d,e} from the expert classification dataset are obtained. Using these classified items {b,c,d,e} as vertices, the connections between vertices {b,c,d,e} are determined based on their classification labels. This process is repeated to construct the following structure: Figure 10 The diagram showing the classification of medical items among the categorized medical items is shown.

[0111] Understandably, expert classification systems, based on therapeutic applications, define related medical items as a single category. Therefore, based on the principle that medical items belonging to the same group are connected, a medical item classification map can be constructed using the classification labels of already categorized medical items.

[0112] See Figure 2 , Figure 2 This application provides a flowchart illustrating a medical item classification method according to an embodiment. Figure 2 As shown, in some embodiments, step S150 of the medical item classification method provided in this application includes:

[0113] Step S210: Determine the edit distance between the plurality of second complete maps and the medical item classification map.

[0114] It is understandable that when the medical item to be classified in the second complete graph is the same as the classified medical item in the medical item classification graph, the edit distance between the second complete graph and the medical item classification graph is determined.

[0115] For example, such as Figure 7 and Figure 8 As shown, the medical items to be classified in the second complete graph G′ are {a,b,c,d,e}, and the medical item classification graph G′ is shown. * Given the categorized medical items {a,b,c,d,e}, determine the second complete graph G′ and the medical item classification graph G. * Edit distance D(G′,G) between * ).

[0116] In one specific embodiment, when the medical item to be classified in the second complete map is different from the classified medical item in the medical item classification map, the second complete map is cropped according to the classified medical item in the medical item classification map, and the edit distance between the cropped second complete map and the medical item classification map is determined.

[0117] For example, such as Figure 7 and Figure 10As shown, the medical items to be classified in the second complete graph G′ are {a,b,c,d,e}, and the medical item classification graph G′ is shown. * If the categorized medical items in the graph are {b,c,d,e}, then based on the categorized medical items {b,c,d,e}, the second complete graph G1' is clipped to obtain the following: Figure 9 The cropped second complete graph G″ is shown. The medical items to be classified in the second complete graph G″ are {b,c,d,e}. Determine whether the cropped second complete graph G″ matches the medical item classification graph G. * Edit distance D(G″,G) between * ).

[0118] It should be noted that the calculation of the edit distance between two images is known to those skilled in the art and will not be elaborated here.

[0119] Step S220: Based on the edit distance, determine the similarity measure between the plurality of second complete maps and the medical item classification map.

[0120] Understandably, multiple edit distances are obtained between the multiple second complete graphs and the medical item classification graph. Based on these multiple edit distances, a similarity measure between the multiple second complete graphs and the medical item classification graph is determined.

[0121] For example, the second complete graph G′ and the medical item classification graph G * The edit distance between them is D(G′,G * If the similarity measure is S(G′,G), then... * )=1 / D(G′,G * ).

[0122] Step S230: Based on the similarity metric, determine the classification label for the medical item to be classified.

[0123] See Figure 3 , Figure 3 This application provides a flowchart illustrating a medical item classification method according to an embodiment. Figure 3 As shown, in some embodiments, step S230 of the medical item classification method provided in this application includes:

[0124] Step S310: Take the maximum value in the similarity measure as the target similarity measure, and determine the complete subgraph in the second complete graph corresponding to the target similarity measure.

[0125] Understandably, based on the definition of edit distance and the relationship between similarity metric and edit distance, the greater the similarity metric between the second complete graph and the medical item classification graph, the more similar the second complete graph is to the medical item classification graph. In other words, the classification results of the medical items to be classified represented by the second complete graph are more similar to the expert classification results.

[0126] For example, such as Figure 7 The second complete graph G′ shown includes a complete subgraph G′. 11 G′ 12 and G′ 13 , where the complete subgraph G′ 11 Including the medical items to be classified {a,b,e}, and the complete subgraph G′ 12 Including the medical items to be classified {b,c,e}, the complete subgraph G′ 13 This includes medical items to be classified {c,d,e}.

[0127] It should be noted that the second complete graph has been cropped, and the medical items to be classified in its complete subgraphs are highly similar to each other, which can be used as the classification result of medical items.

[0128] Step S320: Determine the classification label of the medical item to be classified based on the complete sub-graph.

[0129] It is understandable that unclassified medical items in the complete subgraph are considered to be in the same group, meaning that unclassified medical items in the complete subgraph carry the same classification label.

[0130] For example, the complete subgraph G′ 11 If the medical items to be classified are {a,b,e}, then the medical items to be classified {a,b,e} belong to the same group of medical items and carry the same classification label.

[0131] Based on the second complete graph corresponding to the target similarity measure, a complete subgraph is obtained in the second complete graph. The medical items to be classified in the complete subgraph are regarded as the same group of medical items, and the classification result of the medical items to be classified in the second complete graph is obtained.

[0132] The medical item classification method provided in this application is described below with a specific embodiment:

[0133] Obtain the patient visit dataset, which includes visit data from multiple patients. This dataset includes the medical services {a, b, c, d, e} that each patient visited. Based on this dataset, count the number of patients who used the medical services to be classified, using the formula:

[0134]

[0135] Determine the similarity value D between each pair of medical items to be classified. Based on the similarity values ​​D, construct a similarity matrix A between all medical items to be classified.

[0136]

[0137] Construct a first complete graph based on the similarity matrix A, that is, using the medical items to be classified {a,b,c,d,e} as vertices, and based on the similarity value {D}. ab D ac D ad D ae D bc D bd D be D cd D ce D de Construct weighted edges, and build as follows: Figure 6 The first complete graph G is shown. A similarity threshold D is defined. * Delete the first complete graph G whose weights are less than the similarity threshold D. * The weighted edges are obtained as follows Figure 7 The second complete graph G′ is shown.

[0138] Based on the second complete graph G′, obtain an expert classification dataset including classified medical items with classification labels. Using the classified medical items as vertices, determine the connections between vertices according to the classification labels of the classified medical items, and construct a model as follows: Figure 10 The medical item classification diagram G shown * Based on the categorized medical items in the medical item classification map, the second complete graph G′ is clipped to obtain the following result: Figure 9 The second complete graph G″ is shown.

[0139] Determine the relationship between the cropped second complete image G″ and the medical item classification image G. * Edit distance D(G″,G) between * According to the edit distance D(G″,G), * ), determine the second complete graph G″ and the medical item classification graph G * The similarity measure between them is S(G′,G * )=1 / D(G′,G * Adjust the similarity threshold D * Different second complete graphs are obtained by cropping the first complete graph based on different similarity thresholds.

[0140] Obtain similarity measures based on different second complete graphs, and take the maximum value of the similarity measures as the target similarity measure. Determine the complete subgraph in the second complete graph corresponding to the target similarity measure, regard the medical items to be classified in the complete subgraph as the same group of medical items, obtain the classification result of the medical items to be classified in the second complete graph, and determine the classification label of the medical items to be classified.

[0141] This application proposes a medical item classification method. The method involves acquiring a medical visit dataset containing multiple patient visit records, each including at least one medical item to be classified. Based on the dataset, similarity values ​​are determined between each pair of medical items to be classified. Using each medical item as a vertex, weighted edges are constructed based on the similarity values ​​to build a first complete graph. Then, the first complete graph is pruned based on different similarity thresholds to obtain multiple second complete graphs. The classification labels for the medical items are determined based on the similarity measure between these second complete graphs and a pre-defined medical item classification graph. In this embodiment, the labels for unclassified medical items in the medical visit dataset are supplemented based on the similarity measure and the pre-defined medical item classification graph.

[0142] It is further understood that although the operations are described in a specific order in the accompanying drawings in the embodiments of this application, this should not be construed as requiring these operations to be performed in the specific order or serial order shown, or requiring all the operations shown to be performed to obtain the desired result. In certain environments, multitasking and parallel processing may be advantageous.

[0143] Please see Figure 11 This application embodiment also provides a medical item classification device 100, the medical item classification device 100 comprising:

[0144] The acquisition module 110 is used to acquire a medical visit dataset, which includes medical visit data of multiple patients, and each medical visit dataset includes at least one medical item to be classified for the patient's medical visit.

[0145] The processing module 120 is used to determine the similarity value between each pair of medical items to be classified based on the medical visit dataset.

[0146] The construction module 130 is used to construct a first complete graph between the medical items to be classified, with the medical items to be classified as vertices and weighted edges based on the similarity values ​​between each pair of the medical items to be classified.

[0147] The cropping module 140 is used to crop the first complete image based on different similarity thresholds to obtain multiple second complete images.

[0148] The classification module 150 is used to determine the classification label of the medical item to be classified based on the similarity measure between the plurality of second complete maps and the preset medical item classification map.

[0149] This application proposes a medical item classification device. The device acquires a medical visit dataset containing multiple patient visit data, each data point including at least one medical item to be classified. A processing module determines the similarity value between pairs of medical items based on the dataset. A construction module constructs a first complete graph of the medical items, using each item as a vertex and the similarity value as the vertices. A pruning module then prunes the first complete graph based on different similarity thresholds, resulting in multiple second complete graphs. Finally, a classification module determines the classification labels for the medical items based on a similarity metric between the second complete graphs and a preset medical item classification graph. This embodiment of the application completes the labeling of unclassified medical items in the medical visit dataset based on the similarity metric and the preset medical item classification graph.

[0150] It should be noted that the information interaction and execution process between the modules of the above-mentioned device are based on the same concept as the method embodiment of this application. For details on their specific functions and technical effects, please refer to the method embodiment section, and they will not be repeated here.

[0151] Please see Figure 12 , Figure 12 This application illustrates the hardware structure of an electronic device according to an embodiment of the present application. The electronic device includes:

[0152] The processor 210 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant computer programs to implement the technical solutions provided in the embodiments of this application.

[0153] The memory 220 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). The memory 220 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 220 and called and executed by the processor 210 using the medical item classification method of the embodiments of this application.

[0154] Input / output interface 230 is used to implement information input and output;

[0155] The communication interface 240 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.); and the bus 250 is used to transmit information between each component of the device (such as processor 210, memory 220, input / output interface 230 and communication interface 240).

[0156] The processor 210, memory 220, input / output interface 230 and communication interface 240 are connected to each other within the device via bus 250.

[0157] This application also provides a storage medium, which is a computer-readable storage medium for computer-readable storage. The storage medium stores one or more computer programs, which can be executed by one or more processors to implement the above-mentioned medical item classification method.

[0158] Memory, as a computer-readable storage medium, can be used to store software programs and computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0159] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0160] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0161] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0162] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0163] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0164] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0165] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0166] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0167] Furthermore, the functional units in each embodiment of this application can be integrated into a single processing unit, or each unit can exist independently, or two or more units can be integrated into a single unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0168] If the assembled units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of each embodiment of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0169] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A medical project classification method, characterized in that: include: Acquire a medical consultation data set, wherein the medical consultation data set includes medical consultation data of a plurality of patients, and each medical consultation data includes at least one medical item to be classified during the medical consultation of the patient; Based on the medical consultation data set, determining similarity values ​​between each of the medical items to be classified; Taking the medical items to be classified as vertices, constructing weighted edges according to the similarity values ​​between the medical items to be classified, and constructing a first complete graph between the medical items to be classified; Based on different similarity thresholds, the first complete graph is trimmed to obtain multiple second complete graphs; Determining the classification labels of the medical items to be classified according to the similarity measurement between the plurality of second complete graphs and the preset medical item classification graph specifically includes: determining edit distances between the plurality of second complete graphs and the medical item classification graph; Determining similarity measures between the plurality of second complete graphs and the medical item classification graph according to the edit distance; Based on the similarity measure, a classification label of the medical item to be classified is determined.

2. The method according to claim 1, characterized in that Determining the classification label of the medical item to be classified based on the similarity measurement includes: Taking the maximum value of the similarity measure as the target similarity measure, and determining the complete subgraph in the second complete graph corresponding to the target similarity measure; According to the complete subgraph, a classification label of the medical item to be classified is determined.

3. The medical project classification method according to claim 1, characterized in that: Determining the similarity values ​​between the medical items to be classified based on the medical consultation data set includes: Based on the medical consultation data set, counting the number of patients who have used the medical items to be classified; According to the number of patients, the similarity values ​​between the medical items to be classified are determined.

4. The method according to claim 3, characterized in that The similarity value is determined by the following formula: Among them, the Indicates the medical items to be classified and The similarity value between Indicates that the medical items to be classified have been used The number of patients, Indicates that the medical items to be classified have been used The number of patients, Indicates that the medical items to be classified have been used and of patients.

5. The method according to claim 1, characterized in that The first complete graph is trimmed based on different similarity thresholds to obtain multiple second complete graphs, including: Based on different similarity thresholds, weighted edges with weights less than the similarity threshold in the first complete graph are deleted to obtain multiple second complete graphs.

6. The method according to claim 1, characterized in that After the first complete graph is clipped based on different similarity thresholds to obtain a plurality of second complete graphs, and before the classification labels of the medical items to be classified are determined based on similarity measures between the plurality of second complete graphs and a preset medical item classification graph, the method further includes: According to the second complete graph, obtaining an expert classification dataset, wherein the expert classification dataset includes classified medical items with classification labels; The classified medical items are taken as vertices, connection relationships between the vertices are determined according to classification labels of the classified medical items, and a medical item classification graph between the classified medical items is constructed.

7. A medical project classification device, characterized in that: The device comprises: An acquisition module, used for acquiring a medical consultation data set, wherein the medical consultation data set includes medical consultation data of a plurality of patients, and each medical consultation data includes at least one medical item to be classified during the medical consultation of the patient; A processing module, used for determining the similarity values ​​between the medical items to be classified based on the medical consultation data set; A construction module, used for constructing weighted edges based on similarity values ​​between two medical items to be classified, taking the medical items to be classified as vertices, and constructing a first complete graph between the medical items to be classified; A trimming module, configured to trim the first complete graph based on different similarity thresholds to obtain a plurality of second complete graphs; The classification module is used to determine the classification labels of the medical items to be classified according to the similarity measurement between the plurality of second complete graphs and the preset medical item classification graph, specifically comprising: determining edit distances between the plurality of second complete graphs and the medical item classification graph; Determining similarity measures between the plurality of second complete graphs and the medical item classification graph according to the edit distance; Based on the similarity measure, a classification label of the medical item to be classified is determined.

8. A computer device, characterized in that: include: Memory, used to store programs; A processor, configured to execute the program stored in the memory. When the processor executes the program stored in the memory, the processor is configured to execute: a method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: Computer executable instructions are stored, and the computer executable instructions are used to execute: the method according to any one of claims 1 to 6.

Images