CN110609910B

CN110609910B - Medical knowledge graph construction method and device, storage medium and electronic equipment

Info

Publication number: CN110609910B
Application number: CN201910883707.4A
Authority: CN
Inventors: 陈鹏; 郑彬丽; 陈阳
Original assignee: Golden Panda Ltd
Current assignee: Golden Panda Ltd
Priority date: 2019-09-18
Filing date: 2019-09-18
Publication date: 2023-01-31
Anticipated expiration: 2039-09-18
Also published as: CN110609910A

Abstract

The invention discloses a medical knowledge graph construction method, a medical knowledge graph construction device, a storage medium and electronic equipment, and relates to the technical field of computers. The map construction process related to the medical knowledge map construction method comprises the following steps: acquiring medical data; extracting a plurality of target medical entities from the medical data and determining the relationship among the plurality of target medical entities; if the current medical knowledge graph construction process is not the first medical knowledge graph construction process, fusing the relationships between the multiple target medical entities and the multiple target medical entities into the previous medical knowledge graph to obtain the relationships between the multiple candidate medical entities and the multiple candidate medical entities of the current medical knowledge graph; the current medical knowledge-graph is constructed based on relationships between the plurality of candidate medical entities and the plurality of candidate medical entities of the current medical knowledge-graph. The method and the device improve the construction efficiency of the medical knowledge map.

Description

Medical knowledge graph construction method and device, storage medium and electronic equipment

Technical Field

The present disclosure relates to the field of computer technologies, and in particular, to a medical knowledge graph construction method, a medical knowledge graph construction apparatus, a storage medium, and an electronic device.

Background

The knowledge graph is a formal description framework based on semantic knowledge, and generally nodes are used for representing semantic symbols, and connecting lines between the nodes can be used for representing semantic relationships between the symbols.

Aiming at the medical field, the cost for constructing the medical knowledge map is usually higher due to factors such as strong professional and complex knowledge structure. The manual map construction is not suitable for reality due to high cost, long time consumption and the like. With the development of computer technology, a medical knowledge map can be generated by a machine, however, since the amount of medical information is enormous, there is a problem that the construction efficiency is low when constructing the knowledge map.

It is to be noted that the information disclosed in the above background section is only for enhancement of understanding of the background of the present disclosure, and thus may include information that does not constitute prior art known to those of ordinary skill in the art.

Disclosure of Invention

The present disclosure is directed to a medical knowledge graph construction method, a medical knowledge graph construction apparatus, a storage medium, and an electronic device, which overcome, at least to some extent, the problem of low medical knowledge graph construction efficiency due to limitations and defects of the related art.

According to a first aspect of the present disclosure, there is provided a medical knowledge-graph construction method, including a plurality of medical knowledge-graph construction processes that are iterated a plurality of times, each medical knowledge-graph construction process including: acquiring medical data; extracting a plurality of target medical entities from the medical data and determining the relationship among the plurality of target medical entities; if the current medical knowledge graph construction process is not the first medical knowledge graph construction process, fusing the relationships between the plurality of target medical entities and the plurality of target medical entities into the previous medical knowledge graph to obtain the relationships between the plurality of candidate medical entities and the plurality of candidate medical entities of the current medical knowledge graph; the current medical knowledge-graph is constructed based on relationships between the plurality of candidate medical entities and the plurality of candidate medical entities of the current medical knowledge-graph.

Optionally, constructing the current medical knowledge-graph based on the plurality of candidate medical entities of the current medical knowledge-graph and the relationship between the plurality of candidate medical entities comprises: and acquiring labeling results aiming at the relationships among the candidate medical entities and the candidate medical entities, and constructing the current medical knowledge graph according to the labeling results.

Optionally, the medical knowledge-map construction method further includes: and if the current medical knowledge graph construction process is the first medical knowledge graph construction process, acquiring labeling results aiming at the relationships among the target medical entities and the target medical entities so as to construct the current medical knowledge graph.

Optionally, fusing the plurality of target medical entities and the relationship between the plurality of target medical entities into the previous medical knowledge-graph comprises: obtaining a plurality of historical medical entities in a previous medical knowledge graph and relations among the plurality of historical medical entities; determining a first medical entity of the plurality of target medical entities as a synonym of a target historical entity if a similarity between the first medical entity and a target historical entity of the plurality of historical medical entities is greater than a first similarity threshold; and if a target relationship exists between the first medical entity and a second medical entity in the plurality of target medical entities, and the second medical entity does not exist in the previous medical knowledge-graph, constructing a target relationship between the target historical entity and the second medical entity.

Optionally, the medical knowledge-map construction method further includes: and if the similarity between the third medical entity in the target medical entities and each historical medical entity in the historical medical entities is smaller than a second similarity threshold value, marking the third medical entity, and adding the marked third medical entity to the previous medical knowledge map.

Optionally, the medical data includes one or more of medical standards, medical literature, clinical medical records.

Optionally, the entity determination module is configured not to perform: a plurality of keywords are extracted from the medical data by using a natural language processing model, and the keywords are used as a plurality of target medical entities.

Optionally, the multi-iterative medical knowledge map construction process comprises a plurality of construction processes that are iteratively performed, the construction process being a construction of a knowledge map based on medical knowledge, or a construction of a knowledge map based on real-world clinical data.

Optionally, when the medical knowledge is a medical guideline, the medical knowledge graph constructed based on the medical guideline includes synonyms and superior-inferior relations of the medical entities; when the medical knowledge is a medical standard, a medical knowledge map constructed based on the medical standard comprises standard descriptions and relation descriptions of various medical entities; the knowledge graph constructed based on real world clinical data includes clinical medical entities and entity relationships.

Optionally, extracting the plurality of target medical entities from the medical profile comprises: a natural language processing model is used for extracting a plurality of keywords from the medical data, and the keywords are used as a plurality of target medical entities.

Optionally, after acquiring the medical data, the medical knowledge-map construction method further includes: judging whether the medical data is applied to the construction process of the medical knowledge graph or not based on the generation time and the name of the medical data; discarding the medical data if the medical data has been applied to the construction process of the medical knowledge-graph; if the medical data is not applied to the construction process of the medical knowledge-graph, a process of extracting a plurality of target medical entities from the medical data is performed.

According to a second aspect of the present disclosure, a medical knowledge graph construction apparatus is provided, which includes a data acquisition module, an entity determination module, an entity fusion module, and a first graph construction module.

Specifically, the data acquisition module is used for acquiring medical data; the entity determining module is used for extracting a plurality of target medical entities from the medical data and determining the relationship among the plurality of target medical entities; the entity fusion module is used for fusing the relationships between the plurality of target medical entities and the plurality of target medical entities into the previous medical knowledge map to obtain the relationships between the plurality of candidate medical entities and the plurality of candidate medical entities of the current medical knowledge map if the current medical knowledge map construction process is not the first medical knowledge map construction process; the first atlas construction module is used for constructing the current medical knowledge-atlas based on relationships between a plurality of candidate medical entities and a plurality of candidate medical entities of the current medical knowledge-atlas.

Optionally, the first atlas building module is configured to perform: and acquiring labeling results aiming at the relationships among the candidate medical entities and the candidate medical entities, and constructing the current medical knowledge graph according to the labeling results.

Optionally, the medical knowledge-map construction apparatus further comprises a second map construction module.

In particular, the second atlas construction module is configured to perform: and if the current medical knowledge graph construction process is the first medical knowledge graph construction process, acquiring labeling results aiming at the plurality of target medical entities and the relationships among the plurality of target medical entities so as to construct the current medical knowledge graph.

Optionally, the entity fusion module is configured to perform: obtaining a plurality of historical medical entities in a previous medical knowledge graph and relations among the plurality of historical medical entities; determining a first medical entity of the plurality of target medical entities as a synonym of a target historical entity if a similarity between the first medical entity and a target historical entity of the plurality of historical medical entities is greater than a first similarity threshold; and if a target relationship exists between the first medical entity and a second medical entity in the plurality of target medical entities, and the second medical entity does not exist in the previous medical knowledge-graph, constructing a target relationship between the target historical entity and the second medical entity.

Optionally, the entity fusion module is further configured to perform: and if the similarity between the third medical entity in the target medical entities and each historical medical entity in the historical medical entities is smaller than a second similarity threshold value, marking the third medical entity, and adding the marked third medical entity to the previous medical knowledge graph.

Optionally, the entity determination module is configured to not perform: a plurality of keywords are extracted from the medical data by using a natural language processing model, and the keywords are used as a plurality of target medical entities.

Optionally, the medical knowledge map construction apparatus further comprises a material verification module.

Specifically, the profile verification module is configured to perform: judging whether the medical data is applied to the construction process of the medical knowledge graph or not based on the generation time and the name of the medical data; discarding the medical data if the medical data has been applied to the construction process of the medical knowledge-graph; if the medical data is not applied to the construction process of the medical knowledge-graph, a process of extracting a plurality of target medical entities from the medical data is performed.

According to a third aspect of the present disclosure, there is provided a storage medium having stored thereon a computer program which, when executed by a processor, implements the medical knowledge-map construction method of any one of the above.

According to a fourth aspect of the present disclosure, there is provided an electronic device comprising: a processor; and a memory for storing executable instructions for the processor; wherein the processor is configured to perform any of the above-described medical knowledge-map construction methods via execution of executable instructions.

In some embodiments of the present disclosure, a medical knowledge graph construction method includes a plurality of iterative medical knowledge graph construction processes, where the medical knowledge graph construction processes are sequentially performed, and in each medical knowledge graph construction process, medical data is acquired, a plurality of target medical entities are extracted from the medical data, and relationships between the target medical entities are determined, and if a current medical knowledge graph construction process is not a first medical knowledge graph construction process, relationships between the plurality of target medical entities and the plurality of target medical entities may be fused to a previous medical knowledge graph, and a current medical knowledge graph is determined based on a fused result. On one hand, entities contained in new medical data and the relationship among the entities are fused into the constructed medical knowledge graph, and the construction efficiency of the medical knowledge graph is greatly improved through the continuous iteration processing mode; on the other hand, the process of constructing the medical knowledge graph is dynamic, once new medical data are input, the medical knowledge graph is updated, so that the medical knowledge graph is more accurate and can be better applied to a real scene due to timely matching with the current real situation; in another aspect, the present disclosure employs multiple medical knowledge map building processes, which can store the result of each building process for backtracking, and avoid the problem that the map building process must be restarted due to a data error.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure. It is to be understood that the drawings in the following description are merely exemplary of the disclosure, and that other drawings may be derived from those drawings by one of ordinary skill in the art without the exercise of inventive faculty. In the drawings:

FIG. 1 schematically shows a flow diagram of one medical knowledge-map construction process involved in a medical knowledge-map construction method according to an exemplary embodiment of the present disclosure;

FIG. 2 shows a schematic diagram of a medical data source according to an exemplary embodiment of the present disclosure;

FIG. 3 schematically shows a flow chart of an overall process of constructing a medical knowledge-map according to an exemplary embodiment of the present disclosure;

FIG. 4 shows a schematic diagram of a medical knowledge-map formed, for example, from gastric cancer, according to an exemplary embodiment of the present disclosure;

FIG. 5 schematically shows a block diagram of a medical knowledge-map construction apparatus according to an exemplary embodiment of the present disclosure;

FIG. 6 schematically shows a block diagram of a medical knowledge-map construction apparatus according to another exemplary embodiment of the present disclosure;

FIG. 7 schematically illustrates a block diagram of a medical knowledge-map construction apparatus according to yet another exemplary embodiment of the present disclosure;

fig. 8 schematically illustrates a block diagram of an electronic device according to an exemplary embodiment of the present disclosure.

Detailed Description

Example embodiments will now be described more fully with reference to the accompanying drawings. Example embodiments may, however, be embodied in many different forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of embodiments of the disclosure. One skilled in the relevant art will recognize, however, that the embodiments of the present disclosure can be practiced without one or more of the specific details, or with other methods, components, devices, steps, etc. In other instances, well-known technical solutions have not been shown or described in detail to avoid obscuring aspects of the present disclosure.

Furthermore, the drawings are merely schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus their repetitive description will be omitted. Some of the block diagrams shown in the figures are functional entities and do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different networks and/or processor devices and/or microcontroller devices.

The flow charts shown in the drawings are merely illustrative and do not necessarily include all of the steps. For example, some steps may be decomposed, some steps may be combined or partially combined, and thus the actual execution order may be changed according to the actual situation. In addition, the terms "first," "second," and "third" are used for distinguishing purposes only and should not be construed as limiting the present disclosure.

In the process of constructing the medical knowledge graph, one situation is that construction is completely carried out by depending on experts or professionals, and although the accuracy is guaranteed, the cost is high and the time consumption is long; in another case, the construction of the knowledge map is performed only by a computer, but the amount of medical information is enormous and complicated for a special field such as medicine, and there is a problem that the accuracy is not high and the construction efficiency is low when the medical information is processed in a unified manner.

The medical knowledge map construction method described below may be implemented by a server, that is, the steps involved in the medical knowledge map construction method may be performed by the server, in which case the medical knowledge map construction apparatus of the exemplary embodiment of the present disclosure may be configured within the server.

The medical knowledge-map construction method of the exemplary embodiments of the present disclosure includes a plurality of iterative medical knowledge-map construction processes, which may be sequentially performed. After the construction process of each medical knowledge map is completed, a corresponding medical knowledge map can be obtained, in this case, the tth medical knowledge map can be determined based on the tth-1 medical knowledge map, wherein t is a positive integer greater than 1.

It should be understood that the medical knowledge-map obtained by the present disclosure may be a knowledge-map finally determined through a plurality of construction processes, and then the knowledge-map is applied to reality as a product. In addition, the medical knowledge graph obtained by the method can be a knowledge graph which is continuously and dynamically updated, and can be optimized while being applied to reality.

Fig. 1 schematically shows a flow chart of one medical knowledge-map construction process involved in the medical knowledge-map construction method of an exemplary embodiment of the present disclosure. Referring to fig. 1, a particular method process may include the steps of:

s12, medical data are obtained.

In an exemplary embodiment of the present disclosure, referring to fig. 2, the medical profile may include any profile related to medical information. In particular, medical standards, medical guidelines, medical literature, medical news, etc., may be included, which may relate to medical writings, medical papers, medical reports, etc.

In view of the lack of real world clinical data in the process of constructing the medical knowledge graph at present, the medical data adopted in the exemplary embodiment of the disclosure further include clinical data such as medical orders and medical records, and therefore, the accuracy and the application value of the constructed medical knowledge graph are greatly improved.

In addition, since the medical knowledge graph construction method according to the exemplary embodiment of the present disclosure includes a plurality of medical knowledge graph construction processes, the embodiments of the present disclosure do not limit the number or size of the acquired medical data for each medical knowledge graph construction process, for example, one medical work may be acquired as the medical data, hundreds of medical papers may be acquired as the medical data, thousands of medical records may be acquired as the medical data, and so on.

In order to avoid the problem of wasting system resources due to repeated construction of medical knowledge maps, medical data acquired in the construction process of each medical knowledge map are different. In this case, some embodiments of the present disclosure may further include a process of verifying the acquired medical data.

Whether the acquired medical data is applied to the construction process of the medical knowledge graph or not can be judged based on the generation time and the name of the medical data. Specifically, the server may store the generation time and name of the medical data each time the atlas construction is performed using the medical data. When new medical data is acquired, the generation time and the name of the new medical data can be compared with information stored in the server in advance to determine whether the new medical data is applied to the construction process of the medical knowledge graph.

In addition, in the case that whether the medical data is applied to the atlas construction process cannot be determined by using the generation time and the name, other elements for verification can be added in combination with the source of the medical data. For example, taking a medical record as an example, the name of the patient can be added; as another example, in the case of a medical paper, an author name may be added, and so on.

It should be appreciated that for the case where, for example, thousands of medical records are acquired at a time as medical material, the server determines, on a per-case basis, whether each medical record has been applied to the construction process of the medical knowledge-graph.

When it is determined that the acquired medical data is applied to the construction process of the medical knowledge graph, the server can directly discard the medical data to finish the construction process of the medical knowledge graph.

Upon determining that the acquired medical data is not applied to the construction process of the medical knowledge-graph, the server may perform the following step S14.

S14, extracting a plurality of target medical entities from the medical data, and determining the relationship among the plurality of target medical entities.

In an exemplary embodiment of the present disclosure, a natural language processing model may be used to extract a plurality of keywords from the medical data obtained in step S12, and the plurality of keywords may be used as the target medical entity.

In particular, the natural language processing model referred to herein may include a word segmentation module and a keyword extraction module. The word segmentation module can perform word segmentation processing on the text related to the medical data and stop words. The keyword extraction module may process a result output by the word segmentation module to obtain a plurality of keywords, where the keyword extraction module may be constructed by using TF-IDF (Term Frequency-Inverse text Frequency index), textRank, a keyword extraction algorithm based on semantics, and the like, which is not particularly limited in this exemplary embodiment.

In the case of multiple target medical entities, the relationship between the target medical entities may be determined in conjunction with the context relationship based on semantic analysis means.

In addition, the server can visually display the determined target medical entities and the relationship among the target medical entities so as to be adjusted manually.

According to some embodiments of the disclosure, the server may show the determined keywords and the relationships between the keywords to developers, and the developers may check the keywords to determine the relationships between the plurality of target medical entities and the plurality of target medical entities.

In addition, the relationships among the target medical entities and the target medical entities can be determined through artificial analysis of the medical data.

And S16, if the current medical knowledge graph construction process is not the first medical knowledge graph construction process, fusing the relationships between the plurality of target medical entities and the plurality of target medical entities into the previous medical knowledge graph to obtain the relationships between the plurality of candidate medical entities and the plurality of candidate medical entities of the current medical knowledge graph.

The server may determine whether the currently performed medical knowledge-graph construction process is a first-time medical knowledge-graph construction process. Specifically, if the constructed medical knowledge map does not exist, the current medical knowledge map construction process is not the first medical knowledge map construction process. Or, the server stores an identifier for executing the medical knowledge graph construction times, and whether the current medical knowledge graph construction process is the first medical knowledge graph construction process can be determined through the identifier.

If the current medical knowledge-graph construction process is not the first medical knowledge-graph construction process, the server may fuse the plurality of target medical entities and the relationships between the plurality of target medical entities into a previous medical knowledge-graph.

Specifically, the server may obtain a plurality of historical medical entities and relationships between the historical medical entities in the previous medical knowledge graph, so as to perform a fusion process of the entities and the corresponding relationships. It is easy to understand that the last medical knowledge-graph includes all the medical entities and relationships between the entities determined by the executed medical knowledge-graph construction process.

According to some embodiments of the present disclosure, the server may determine a similarity between each target medical entity and the historical entities. Specifically, the similarity between the entities may be determined by a method for measuring the similarity of the character strings, for example, the similarity may be measured by using TF-IDF and Jaccard (Jaccard), which is not limited in this exemplary embodiment.

A first medical entity of the plurality of target medical entities may be determined to be a synonym for a target historical entity of the plurality of historical medical entities if a similarity between the first medical entity and the target historical entity is greater than a first similarity threshold. The present disclosure does not particularly limit the specific numerical value of the first similarity threshold.

If, among a plurality of target medical entities determined based on the acquired medical data, the first medical entity and the second medical entity have a relationship (denoted as a target relationship), and it is determined that the second medical entity does not exist in the previous medical knowledge-graph, or it is determined that no relationship exists between the second medical entity and each historical entity, a target relationship between the target historical entity and the second medical entity can be constructed.

For example, the first medical entity in the target medical entities is "coronary atherosclerotic heart disease", the second medical entity is "angina pectoris", and the relationship between the first medical entity and the second medical entity is determined as follows: the second medical entity is a symptom of the first medical entity. Among historical medical entities there is a target historical entity, "coronary heart disease". In this case, "coronary atherosclerotic heart disease" may be regarded as a synonym for "coronary heart disease", and a symptom relationship of "coronary heart disease" and "angina pectoris" is constructed when "angina pectoris" is not present in the last medical knowledge map.

According to further embodiments of the present disclosure, if the similarity between a third medical entity of the target medical entities and each historical medical entity of the historical medical entities is less than the second similarity threshold (which may be set by human beings and is less than the first similarity threshold), it indicates that the third medical entity is not in great relationship with each historical medical entity, in which case, the third medical entity may be marked, and the marked third medical entity may be added to the previous medical knowledge-graph. So that the third medical entity can be determined quickly in the following and analyzed, and whether the position of the third medical entity in the knowledge graph and the corresponding relation are accurate or not can be determined conveniently.

After fusing the plurality of target medical entities and the relationships between the plurality of target medical entities into a previous medical knowledge-graph, a plurality of candidate medical entities of the current medical knowledge-graph constructed by the current medical knowledge-graph construction process and the relationships between the candidate medical entities may be determined.

And S18, constructing the current medical knowledge graph based on the relationships among the candidate medical entities and the candidate medical entities of the current medical knowledge graph.

According to some embodiments of the present disclosure, the current medical knowledge-graph is constructed directly using relationships between the plurality of candidate medical entities and the plurality of candidate medical entities. That is, the plurality of candidate medical entities are entities in the determined current medical knowledge-graph, and the relationship between the plurality of candidate medical entities is the relationship between the entities in the determined current medical knowledge-graph.

According to other embodiments of the present disclosure, the server may obtain labeling results for the plurality of candidate medical entities and the relationships between the plurality of candidate medical entities, and construct the current medical knowledge-graph according to the labeling results. Wherein, the annotation may be understood as modifying, deleting, supplementing, etc. the plurality of candidate medical entities generated by the server and the relationship between the plurality of candidate medical entities. The labeling process may be a manual operation, however, in some instances where machine learning is applied, labeling may also be implemented by a machine learning model, which is not limited in this exemplary embodiment.

For the timing of labeling, in one embodiment, there may be a process of labeling relationships between a plurality of candidate medical entities and a plurality of candidate medical entities in each medical knowledge graph construction process.

In another embodiment, the process of labeling may be performed at predetermined time intervals (e.g., one day, one week), in which case the medical knowledge-graph construction process may have been performed several times.

In yet another embodiment, the process of labeling may be performed after a predetermined number of atlas creation processes are performed. The predetermined number of times may be set manually, and the specific value is not limited in the embodiment of the present disclosure.

The above description is directed to the case where the current medical knowledge-graph construction process is not the first medical knowledge-graph construction process. In contrast, in the case that the current medical knowledge-graph constructing process is the first medical knowledge-graph constructing process, after step S14, the medical knowledge-graph constructing method according to the exemplary embodiment of the present disclosure further includes: and constructing the current medical knowledge graph by using the relationships among the plurality of target medical entities and the plurality of target medical entities.

According to one embodiment of the present disclosure, the plurality of target medical entities are entities of the determined current medical knowledge-graph, and the relationship between the plurality of target medical entities is the relationship between the entities of the determined current medical knowledge-graph.

According to another embodiment of the present disclosure, the server may obtain labeling results for a plurality of target medical entities and a relationship between the plurality of target medical entities, and construct the current medical knowledge-graph according to the labeling results.

The entire process of constructing a medical knowledge graph spectrum according to an exemplary embodiment of the present disclosure will be described below with reference to fig. 3.

In step S302, the server may obtain medical data, which may include one or more of medical standards, medical literature, and clinical medical records.

In step S304, the server may determine whether the medical data has been used in the construction process of the medical knowledge-graph, and if so, perform step S306; if not, step S308 is performed. In addition, if the medical data includes a part that is used, the used part is executed in step S306, and the unused part is used to execute step S308.

In step S306, the server may discard the medical data or perform no processing on the medical data.

In step S308, the server may extract the target medical entities from the medical data using the natural language processing model and determine the relationships between the target medical entities using the context relationships.

In step S310, it is determined whether the currently performed medical knowledge-map construction process is a first-time medical knowledge-map construction process. If yes, go to step S312; if not, step S314 is performed.

In step S312, the multiple target medical entities and the relationships between the multiple target medical entities are manually labeled.

In step S314, the server may fuse the plurality of target medical entities and the relationships between the plurality of target medical entities into a previous medical knowledge-graph. In step S316, the fused medical knowledge graph is manually labeled.

In step S318, a current medical knowledge-map may be obtained.

It should be understood that when new medical data is entered, the current medical knowledge-map may be used as the previous medical knowledge-map for determining the medical knowledge-map after the new medical data is merged. That is, step S318 may jump to step S302 to perform another new medical knowledge-map construction process.

Fig. 4 schematically shows a part of the content of one medical knowledge-map determined in performing a plurality of medical knowledge-map building processes according to an exemplary embodiment of the present disclosure.

Referring to fig. 4, it can be determined that synonyms of "gastric cancer" include "gastric cancer", "gastric malignant tumor", "the site of onset of gastric cancer" is "stomach", the international disease with symptoms of "nausea, vomiting", "gastric cancer" is classified as "gastric malignant tumor", the medical guideline corresponds to "gastric malignant tumor", and "gastric cancer" is "gastric cancer" in the international medical dictionary. It is easily understood that "synonyms", "disease sites", "symptoms", "international disease classification (ICD)", "medical guidelines", "national medical phrase dictionary (MedDRA)" are relationships between respective entities.

It should be noted that fig. 4 is only an exemplary depiction, and for the entity of "gastric cancer", the corresponding entity may further include some treatment means, medication, authoritative doctor, hospital, etc., which is not limited in this exemplary embodiment.

In some embodiments of the present disclosure, the multi-iteration medical knowledge-graph construction process may include multiple construction processes, where the construction process may be the construction of a knowledge-graph based on medical knowledge, or the construction process may be the construction of a knowledge-graph based on real-world clinical data.

The medical knowledge may include medical guidelines, medical standards, and the like. The medical criteria may, for example, include international disease classification criteria, national medical phrase dictionary criteria, and the like.

When the medical knowledge is medical guidelines, the medical knowledge graph constructed based on the medical guidelines may include synonyms and superior-inferior relationships of the medical entities.

When medical knowledge is a medical standard, a medical knowledge-graph constructed based on the medical standard may include standard descriptions and relationship descriptions of various medical entities.

In addition, the knowledge graph constructed based on real-world clinical data comprises clinical medical entities and entity relationships.

In some embodiments of the present disclosure, the multiple iterative medical knowledge map construction process includes iteratively performed sets of construction processes, each set of construction processes may include a first construction process and a second construction process.

In the first construction process, the medical data is medical knowledge, and the medical knowledge may include medical guidelines and medical standards. In this case, the first building process may include a first building sub-process and a second building sub-process.

In the first construction subprocess, the vocabulary of each medical field can be determined from the medical guide, the synonym and the superior-inferior relation corresponding to each medical field can be determined from the vocabulary, and a basic medical knowledge map with relative independence of each medical field can be constructed. Referring to fig. 4, medical guidelines for "gastric cancer", "gastric cancer" refer to entities and some synonyms related to "gastric cancer" generated during this process.

In the second construction sub-process, medical standards, such as international disease classification standards, national medical phrase dictionary standards, and the like, may be introduced. Vocabulary is determined from the medical standards to complete the medical knowledge-graph generated by the first construction sub-process.

In a second construction process of medical knowledge opposition, the medical data can include real world clinical data, e.g., medical record data. From the clinical data, terms of various medical fields and diagnostic information, such as anatomical regions, symptoms, etc., can be determined. These entities and corresponding relationships determined based on the clinical data may then be incorporated into the medical knowledge-graph generated by the first construction process described above.

By iteratively executing the above-mentioned group construction processes, the content of the medical atlas is enriched. It should be understood that the above medical guidelines, medical standards, etc. do not perform information extraction once, but mine a part of data each time, and then fuse the mined data into the generated knowledge graph, so that the knowledge graph can be enriched continuously. In addition, knowledge in the knowledge graph can be adopted to assist knowledge mining so as to obtain more entities and corresponding relations.

It should be noted that, within a set of building processes, the present disclosure does not limit the order of the first building process and the second building process, that is, the second building process may be executed first, and then the first building process may be executed. Similarly, the present disclosure does not limit the order of the first and second building sub-processes in the first building process. It is to be readily understood that it is also within the contemplation of the present disclosure that the first build sub-process is performed, followed by the second build sub-process, and then the second build sub-process.

The constructed knowledge graph can comprise synonym relations of medical entities, upper and lower relations of each medical entity, relations among different medical fields, diagnosis and treatment information and the like.

The medical knowledge graph constructed by the construction method of the medical knowledge graph can be efficiently searched, for example, the complications, treatment means, medication, cure rate and the like of a disease can be determined. Has wide application scenes in a plurality of medical fields such as medical theory research, drug development, clinical treatment and the like.

It should be noted that although the various steps of the methods of the present disclosure are depicted in the drawings in a particular order, this does not require or imply that these steps must be performed in this particular order, or that all of the depicted steps must be performed, to achieve desirable results. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step execution, and/or one step broken down into multiple step executions, etc.

Further, a medical knowledge-map constructing apparatus is also provided in the present exemplary embodiment.

Fig. 5 schematically shows a block diagram of a medical knowledge-map construction apparatus of an exemplary embodiment of the present disclosure. Referring to fig. 5, the medical knowledge map construction apparatus 5 according to an exemplary embodiment of the present disclosure may include a profile acquisition module 51, an entity determination module 53, an entity fusion module 55, and a first map construction module 57.

Specifically, the data obtaining module 51 may be configured to obtain medical data; the entity determining module 53 may be configured to extract a plurality of target medical entities from the medical data and determine relationships between the plurality of target medical entities; the entity fusion module 55 may be configured to fuse the plurality of target medical entities and the relationship between the plurality of target medical entities into a previous medical knowledge-graph if the current medical knowledge-graph construction process is not the first medical knowledge-graph construction process, to obtain a plurality of candidate medical entities of the current medical knowledge-graph and a relationship between the plurality of candidate medical entities; the first atlas-constructing module 57 may be used to construct the current medical knowledge-atlas based on relationships between a plurality of candidate medical entities and a plurality of candidate medical entities of the current medical knowledge-atlas.

On the one hand, the medical knowledge graph construction device based on the disclosed exemplary embodiment fuses entities contained in new medical data and relationships among the entities into the constructed medical knowledge graph, and greatly improves the construction efficiency of the medical knowledge graph through the mode of continuous iteration processing; on the other hand, the process of constructing the medical knowledge map is dynamic, once new medical data are input, the medical knowledge map is updated, so that the medical knowledge map is more accurate and can be better applied to a real scene due to timely matching with the current real situation; in another aspect, the present disclosure employs multiple medical knowledge map building processes, which can store the result of each building process for backtracking, and avoid the problem that the map building process must be restarted due to a data error.

According to an exemplary embodiment of the present disclosure, the first atlas construction module 57 may be configured to perform: and acquiring annotation results aiming at the relationships among the candidate medical entities and the candidate medical entities, and constructing the current medical knowledge graph according to the annotation results.

According to an exemplary embodiment of the present disclosure, referring to fig. 6, the medical knowledge-map constructing apparatus 6 may further include a second map constructing module 61, compared to the medical knowledge-map constructing apparatus 5.

In particular, the second atlas construction module 61 may be configured to perform: and if the current medical knowledge graph construction process is the first medical knowledge graph construction process, acquiring labeling results aiming at the relationships among the target medical entities and the target medical entities so as to construct the current medical knowledge graph.

According to an example embodiment of the present disclosure, the entity fusion module 55 may be configured to perform: obtaining a plurality of historical medical entities in a previous medical knowledge graph and relations among the plurality of historical medical entities; determining a first medical entity of the plurality of target medical entities as a synonym of a target historical entity if a similarity between the first medical entity and a target historical entity of the plurality of historical medical entities is greater than a first similarity threshold; and if a target relationship exists between the first medical entity and a second medical entity in the plurality of target medical entities, and the second medical entity does not exist in the previous medical knowledge-graph, constructing a target relationship between the target historical entity and the second medical entity.

According to an exemplary embodiment of the present disclosure, the entity fusion module 55 may be further configured to perform: and if the similarity between the third medical entity in the target medical entities and each historical medical entity in the historical medical entities is smaller than a second similarity threshold value, marking the third medical entity, and adding the marked third medical entity to the previous medical knowledge graph.

According to an exemplary embodiment of the present disclosure, the medical data includes one or more of medical standards, medical literature, clinical medical records.

According to an exemplary embodiment of the present disclosure, the entity determination module 53 may be configured not to perform: a natural language processing model is used for extracting a plurality of keywords from the medical data, and the keywords are used as a plurality of target medical entities.

According to an exemplary embodiment of the present disclosure, the multi-iterative medical knowledge-graph construction process comprises a plurality of construction processes that are iteratively performed, the construction processes being construction of a knowledge-graph based on medical knowledge or construction of a knowledge-graph based on real-world clinical data.

According to the exemplary embodiment of the disclosure, when the medical knowledge is medical guidelines, the medical knowledge graph constructed based on the medical guidelines includes synonyms and superior-inferior relations of the medical entities; when the medical knowledge is a medical standard, a medical knowledge map constructed based on the medical standard comprises standard descriptions and relation descriptions of various medical entities; the knowledge graph constructed based on real world clinical data includes clinical medical entities and entity relationships.

According to an exemplary embodiment of the present disclosure, the medical knowledge-graph construction process of a plurality of iterations comprises a plurality of sets of construction processes performed iteratively, each set of construction processes comprising a first construction process and a second construction process; in the first construction process, the medical data is medical knowledge, and the medical knowledge map constructed in the first construction process comprises one or more of synonyms, superior-inferior relations and medical standard descriptions of all medical field entities; in the second construction process, the medical data is real-world clinical data, and the medical knowledge map constructed by the second construction process comprises diagnosis descriptions of the entities in each medical field.

According to an exemplary embodiment of the present disclosure, the first build process includes a first build sub-process and a second build sub-process; in the first construction subprocess, the medical data is medical guidelines, and the medical knowledge graph constructed in the first construction subprocess comprises synonyms and superior-inferior relations of entities in various medical fields; in the second construction sub-process, the medical data is medical standard, and the medical knowledge graph constructed by the second construction sub-process comprises standard description of each medical field entity.

According to an exemplary embodiment of the present disclosure, referring to fig. 7, the medical knowledge-map construction apparatus 7 may further include a material verification module 71, as compared to the medical knowledge-map construction apparatus 5.

In particular, the material verification module 71 may be configured to perform: judging whether the medical data is applied to the construction process of the medical knowledge graph or not based on the generation time and the name of the medical data; discarding the medical data if the medical data has been applied to the construction process of the medical knowledge-graph; if the medical data is not applied to the construction process of the medical knowledge-graph, a process of extracting a plurality of target medical entities from the medical data is performed.

Since each functional module of the medical knowledge map construction device of the embodiment of the present invention is the same as that of the method and the embodiment of the present invention, the details are not repeated herein.

In an exemplary embodiment of the present disclosure, there is also provided a computer readable storage medium having stored thereon a program product capable of implementing the above-described method of the present specification. In some possible embodiments, the various aspects of the invention may also be implemented in the form of a program product comprising program code means for causing a terminal device to carry out the steps according to various exemplary embodiments of the invention described in the above section "exemplary method" of this description, when said program product is run on said terminal device.

The program product for implementing the above method according to an embodiment of the present invention may employ a portable compact disc read only memory (CD-ROM) and include program codes, and may be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited in this regard and, in the present document, a readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical disk, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

A computer readable signal medium may include a propagated data signal with readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A readable signal medium may also be any readable medium that is not a readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

Program code embodied on a readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server. In situations involving remote computing devices, the remote computing devices may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to external computing devices (e.g., through the internet using an internet service provider).

In an exemplary embodiment of the present disclosure, an electronic device capable of implementing the above method is also provided.

As will be appreciated by one skilled in the art, aspects of the present invention may be embodied as a system, method or program product. Thus, various aspects of the invention may be embodied in the form of: an entirely hardware embodiment, an entirely software embodiment (including firmware, microcode, etc.), or an embodiment combining hardware and software aspects that may all generally be referred to herein as a "circuit," module "or" system.

An electronic device 800 according to this embodiment of the invention is described below with reference to fig. 8. The electronic device 800 shown in fig. 8 is only an example and should not bring any limitations to the function and scope of use of the embodiments of the present invention.

As shown in fig. 8, the electronic device 800 is in the form of a general purpose computing device. The components of the electronic device 800 may include, but are not limited to: the at least one processing unit 810, the at least one memory unit 820, a bus 830 connecting different system components (including the memory unit 820 and the processing unit 810), and a display unit 840.

Wherein the storage unit stores program code that can be executed by the processing unit 810, such that the processing unit 810 performs the steps according to various exemplary embodiments of the present invention described in the above section "exemplary method" of this specification. For example, the processing unit 810 may perform steps S12 to S18 as shown in fig. 1.

The memory unit 820 may include readable media in the form of volatile memory units, such as a random access memory unit (RAM) 8201 and/or a cache memory unit 8202, and may further include a read only memory unit (ROM) 8203.

The storage unit 820 may also include a program/utility 8204 having a set (at least one) of program modules 8205, such program modules 8205 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may comprise an implementation of a network environment.

Bus 830 may be any of several types of bus structures including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

The electronic device 800 may also communicate with one or more external devices 900 (e.g., keyboard, pointing device, bluetooth device, etc.), with one or more devices that enable a user to interact with the electronic device 800, and/or with any device (e.g., router, modem, etc.) that enables the electronic device 800 to communicate with one or more other computing devices. Such communication may occur over input/output (I/O) interfaces 850. Also, the electronic device 800 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network such as the internet) via the network adapter 860. As shown, the network adapter 860 communicates with the other modules of the electronic device 800 via the bus 830. It should be appreciated that although not shown, other hardware and/or software modules may be used in conjunction with the electronic device 800, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.

Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein may be implemented by software, or by software in combination with necessary hardware. Therefore, the technical solution according to the embodiments of the present disclosure may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (which may be a CD-ROM, a usb disk, a removable hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which may be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the embodiments of the present disclosure.

Furthermore, the above-described figures are merely schematic illustrations of processes involved in methods according to exemplary embodiments of the invention, and are not intended to be limiting. It will be readily appreciated that the processes illustrated in the above figures are not intended to indicate or limit the temporal order of the processes. In addition, it is also readily understood that these processes may be performed synchronously or asynchronously, e.g., in multiple modules.

It should be noted that although in the above detailed description several modules or units of the device for action execution are mentioned, such a division is not mandatory. Indeed, the features and functionality of two or more modules or units described above may be embodied in one module or unit, according to embodiments of the present disclosure. Conversely, the features and functions of one module or unit described above may be further divided into embodiments by a plurality of modules or units.

Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice in the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

It will be understood that the present disclosure is not limited to the precise arrangements that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is to be limited only by the terms of the appended claims.

Claims

1. A medical knowledge graph construction method is characterized by comprising a medical knowledge graph construction process of multiple iterations, wherein the result of each construction process is stored, and the medical knowledge graph construction process comprises the following steps:

acquiring medical data; the medical data acquired in the construction process of each medical knowledge map are different;

judging whether the medical data is applied to a construction process of a medical knowledge graph or not based on the generation time and the name of the medical data; and, in case it cannot be determined whether the medical data has been applied to the medical knowledge graph construction process based on the generation time and name of the medical data, determining whether the medical data has been applied to the medical knowledge graph construction process based on the source of the medical data and a preset verification factor;

if the medical data is applied to the construction process of the medical knowledge graph, discarding the medical data to finish the construction process of the medical knowledge graph;

if the medical data is not applied to the construction process of the medical knowledge graph, extracting a plurality of keywords from the medical data, using the keywords as target medical entities, and determining the relationship among the target medical entities based on semantic analysis;

determining that the current medical knowledge graph construction process is not the first medical knowledge graph construction process according to the identification of the times of executing medical knowledge graph construction stored in the server, and fusing the relationships between the plurality of target medical entities and the plurality of target medical entities into the previous medical knowledge graph to obtain a plurality of candidate medical entities of the current medical knowledge graph and the relationships between the plurality of candidate medical entities;

constructing the current medical knowledge-graph based on a plurality of candidate medical entities of the current medical knowledge-graph and relationships between the plurality of candidate medical entities.

2. The medical knowledge-graph construction method according to claim 1, wherein constructing the current medical knowledge-graph based on a plurality of candidate medical entities of the current medical knowledge-graph and relationships between the plurality of candidate medical entities comprises:

and acquiring labeling results aiming at the plurality of candidate medical entities and the relationship among the plurality of candidate medical entities, and constructing the current medical knowledge graph according to the labeling results.

3. The medical knowledge-graph construction method according to claim 1, further comprising:

if the current medical knowledge-graph construction process is a first-time medical knowledge-graph construction process, obtaining labeling results aiming at the target medical entities and relations among the target medical entities so as to construct the current medical knowledge-graph.

4. The medical knowledge-graph construction method of claim 1, wherein fusing the plurality of target medical entities and the relationships between the plurality of target medical entities into a previous medical knowledge-graph comprises:

obtaining a plurality of historical medical entities in the previous medical knowledge-graph and relationships between the plurality of historical medical entities;

determining a first medical entity of the plurality of target medical entities as a synonym of a target historical entity of the plurality of historical medical entities if a similarity between the first medical entity and the target historical entity is greater than a first similarity threshold; and

constructing a target relationship between the target historical entity and a second medical entity of the plurality of target medical entities if the target relationship exists between the first medical entity and the second medical entity does not exist in the previous medical knowledge-graph.

5. The medical knowledge-graph construction method according to claim 2, further comprising:

if the similarity between a third medical entity in the target medical entities and each historical medical entity in the historical medical entities is smaller than a second similarity threshold, marking the third medical entity, and adding the marked third medical entity to the previous medical knowledge map.

6. The medical knowledge graph construction method according to any one of claims 1 to 5, wherein the plurality of iterative medical knowledge graph construction processes comprise a plurality of construction processes which are iteratively executed, and the construction processes are construction of a knowledge graph based on medical knowledge or construction of a knowledge graph based on real world clinical data.

7. The medical knowledge-graph construction method according to claim 6,

when the medical knowledge is a medical guideline, a medical knowledge graph constructed based on the medical guideline comprises synonyms and superior-inferior relations of all medical entities;

when the medical knowledge is a medical standard, a medical knowledge map constructed based on the medical standard comprises standard description and relation description of each medical entity;

the knowledge graph constructed based on real world clinical data includes clinical medical entities and entity relationships.

8. A medical knowledge graph construction apparatus, wherein the apparatus is used for a plurality of iterative medical knowledge graph construction processes, and the result of each construction process is stored, the apparatus comprising:

the data acquisition module is used for acquiring medical data; the medical data acquired in the construction process of each medical knowledge map are different;

the data checking module is used for judging whether the medical data is applied to the construction process of the medical knowledge graph or not based on the generation time and the name of the medical data; and, in case it cannot be judged whether the medical data has been applied to the medical knowledge graph construction process based on the generation time and name of the medical data, judging whether the medical data has been applied to the medical knowledge graph construction process based on the source of the medical data and a preset verification factor; if the medical data is applied to the construction process of the medical knowledge graph, discarding the medical data to finish the construction process of the medical knowledge graph; performing a process of extracting a plurality of target medical entities from medical data if the medical data is not applied to a construction process of a medical knowledge-graph;

the entity determining module is used for extracting a plurality of keywords from the medical data, using the keywords as target medical entities and determining the relation among the target medical entities based on semantic analysis;

an entity fusion module, configured to determine that a current medical knowledge graph construction process is not a first medical knowledge graph construction process according to an identifier of times of performing medical knowledge graph construction stored in a server, and fuse relationships between the plurality of target medical entities and the plurality of target medical entities to a previous medical knowledge graph to obtain a plurality of candidate medical entities of the current medical knowledge graph and relationships between the plurality of candidate medical entities;

a first graph construction module to construct the current medical knowledge-graph based on a plurality of candidate medical entities of the current medical knowledge-graph and relationships between the plurality of candidate medical entities.

9. A storage medium having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the medical knowledge-map construction method of any one of claims 1 to 7.

10. An electronic device, comprising:

a processor; and

a memory for storing executable instructions of the processor;

wherein the processor is configured to perform the medical knowledge-graph construction method of any one of claims 1 to 7 via execution of the executable instructions.