Neurology nursing information intelligent management method

Abstract

The invention relates to the field of nursing information management, in particular to an intelligent management method for nursing information of the neurology department. The method comprises the following steps: firstly, acquiring an electronic medical record of each historical patient in the neurology department, nursing schemes of different types and an electronic medical record of a patient to be detected, and performing word segmentation processing on each electronic medical record and each nursing scheme to obtain segmented words of the electronic medical records and the nursing schemes; based on the content information difference of the same type of nursing schemes among the historical patients, clustering all the historical patients, and according to the occurrence frequency of the segmented words of the electronic medical record of the to-be-detected patient in the electronic medical record of each historical patient in each cluster and the number of the same segmented words, determining the number of the segmented words of the to-be-detected patient in the electronic medical record of each historical patient in each cluster; and obtaining an association degree between the to-be-tested patient and each cluster, and recommending the nursing schemes of the historical patients in the clusters to the to-be-tested patient. According to the invention, the nursing scheme of the to-be-tested patient can be recommended in a targeted and personalized manner, and the effect of neurology nursing information management is improved.

Description

Technical Field

[0001] The present invention relates to the field of nursing information management, and in particular to an intelligent management method for nursing information in neurology. Background Art

[0002] Neurological nursing is professional nursing work for patients with neurological diseases. In the process of nursing neurological patients, targeted and personalized recommendations for patients' care plans are made, which helps to improve treatment effects, enhance patient satisfaction, optimize resource utilization and promote patient recovery.

[0003] In the related art, nursing plans for other patients with similar conditions or symptoms to the patient to be tested are usually recommended to the patient to be tested. However, due to differences in physical constitution and living habits among different patients, even patients with similar conditions or symptoms may require different nursing plans. As a result, existing algorithms cannot provide targeted and personalized recommendations for nursing plans for the patient to be tested, thereby reducing the effectiveness of nursing information management. Summary of the invention

[0004] In order to solve the technical problem that the existing algorithm cannot make targeted and personalized recommendations for the nursing plan of the patient to be tested, and the effect of nursing information management is reduced, the purpose of the present invention is to provide an intelligent management method for nursing information of neurology, and the technical scheme adopted is as follows: The present invention proposes a method for intelligent management of neurology nursing information, the method comprising: Obtain the electronic medical records and different types of nursing plans of each historical patient in the Department of Neurology and the electronic medical records of the patient to be tested, perform word segmentation processing on the electronic medical records and nursing plans of the patient to be tested and each historical patient, and obtain the word segmentation of the electronic medical records and nursing plans; Taking any type of nursing plan of any historical patient as the target nursing plan, performing sentiment analysis on the target nursing plan, and obtaining the sentiment score value of each segmentation in the target nursing plan; performing keyword extraction on the target nursing plan, and obtaining the keywords of the target nursing plan; obtaining the descriptive feature value of each keyword of the target nursing plan according to the sentiment score value of each segmentation in the preset neighborhood of each keyword in the target nursing plan; obtaining the difference in nursing information between any two historical patients according to the difference in the descriptive feature value of the same keyword between the same type of nursing plans of any two historical patients and the number of the same keywords; clustering all historical patients based on the difference in nursing information to obtain multiple clustering clusters; Taking any cluster as the target cluster, obtaining the degree of association between the patient to be tested and the target cluster according to the number of times each word of the electronic medical record of the patient to be tested appears in the electronic medical record of each historical patient in the target cluster, and the number of identical words in the electronic medical records of the patient to be tested and each historical patient in the target cluster; Based on the degree of association between the patient to be tested and each cluster, the nursing plans of the historical patients in each cluster are recommended to the patient to be tested according to different priorities.

[0005] Furthermore, the descriptive feature value of each keyword of the target nursing plan is obtained, including: The average of the sentiment score values ​​of all the segmented words in the preset neighborhood of each keyword in the target nursing plan is used as the descriptive feature value of each keyword in the target nursing plan.

[0006] Furthermore, obtaining the difference in nursing information between any two historical patients includes: Take any type as the target type, and take the set of all keywords of the nursing plan of the target type of each historical patient as the keyword set of each historical patient about the target type; Randomly select two patients as a first historical patient and a second historical patient, perform negative correlation mapping on the Jaccard correlation coefficient between the keyword set of the first historical patient on the target type and the keyword set of the second historical patient on the target type, and obtain a first information difference coefficient between the first historical patient and the second historical patient on the target type; The average of the absolute values ​​of the differences of the descriptive feature values ​​of all the same keywords between the nursing plan of the target type of the first historical patient and the nursing plan of the target type of the second historical patient is taken as the second information difference coefficient about the target type between the first historical patient and the second historical patient; The first information difference coefficient and the second information difference coefficient are integrated to obtain a comprehensive information difference coefficient about the target type between the first historical patient and the second historical patient; The average value of the comprehensive information difference coefficients of all types between the first historical patient and the second historical patient is taken as the nursing information difference degree between the first historical patient and the second historical patient.

[0007] Furthermore, obtaining a plurality of clusters includes: The difference in nursing information between any two historical patients is used as the distance metric between any two historical patients, and a K-means clustering algorithm is used to cluster all historical patients based on the distance metric between any two historical patients to obtain multiple clusters.

[0008] Furthermore, obtaining the degree of association between the patient to be tested and the target cluster includes: Taking any participle of the electronic medical record of the patient to be tested as the target participle, selecting a reference historical patient of the target participle from all historical patients in the target cluster, wherein the target participle appears at least once in the electronic medical record of the reference historical patient; According to the number of reference historical patients of the target segmentation word and the number of times the target segmentation word appears in the electronic medical record of each reference historical patient, the comprehensive distribution breadth of the target segmentation word with respect to the target cluster is obtained; the average value of the comprehensive distribution breadth of all segmentations of the electronic medical record of the patient to be tested with respect to the target cluster is taken as the overall distribution breadth of the patient to be tested with respect to the target cluster; The set consisting of all the segmentations of the electronic medical records of the patient to be tested is used as the first segmentation set of the patient to be tested, any historical patient in the target cluster is used as the target historical patient, and the set consisting of all the segmentations of the electronic medical records of the target historical patient is used as the second segmentation set of the target historical patient; The Jaccard correlation coefficient between the first segmentation set of the patient to be tested and the second segmentation set of the target historical patient is used as the symptom similarity between the patient to be tested and the target historical patient; the average value of the symptom similarities between the patient to be tested and all historical patients in the target cluster is used as the overall similarity between the patient to be tested and the target cluster; The overall similarity between the patient to be tested and the target cluster and the overall distribution breadth of the patient to be tested with respect to the target cluster are integrated and normalized to obtain the degree of association between the patient to be tested and the target cluster.

[0009] Furthermore, obtaining the comprehensive distribution breadth of the target word segmentation with respect to the target cluster includes: The number of all reference historical patients of the target participle is used as the numerator, the number of all historical patients in the target cluster is used as the denominator, and the ratio is used as the first distribution breadth of the target participle in the target cluster; The average number of times the target word appears in the electronic medical records of all reference historical patients is taken as the second distribution breadth of the target word in the target cluster; The first distribution breadth and the second distribution breadth are integrated to obtain a comprehensive distribution breadth of the target word segment with respect to the target cluster.

[0010] Furthermore, the step of recommending the nursing plans of the historical patients in each cluster to the patient to be tested according to different priorities includes: According to the order of the degree of association between the patient to be tested and each cluster from large to small, the nursing plans of the historical patients in each cluster are recommended to the patient to be tested.

[0011] Furthermore, the segmentation of obtaining the electronic medical record and the nursing plan includes: Jieba word segmentation tool is used to perform word segmentation on the electronic medical records and nursing plans of the patients to be tested and historical patients to obtain the word segmentation of the electronic medical records and nursing plans.

[0012] Further, obtaining the sentiment score value of each participle in the target nursing plan includes: The target nursing plan is processed using the NTUSD sentiment dictionary to obtain the sentiment score value of each word in the target nursing plan.

[0013] Furthermore, the keywords for obtaining the target nursing plan include: Using the knowledge word bank, professional vocabulary is extracted from the target nursing plan as the keywords of the target nursing plan.

[0014] The present invention has the following beneficial effects: The present invention takes into account that the existing algorithms cannot make targeted and personalized recommendations for the nursing plans of the patients to be tested, which reduces the effect of nursing information management. Therefore, the electronic medical records of each historical patient in the Department of Neurology and different types of nursing plans and the electronic medical records of the patients to be tested are first obtained, and each electronic medical record and nursing plan is segmented to obtain the segmentation of the electronic medical records and the nursing plans. Subsequently, the differences in electronic medical records and nursing plans between patients can be accurately analyzed based on the segmentation. Considering that different patients have different physical constitutions and living habits, even patients with similar conditions or symptoms may have certain differences in the nursing plans they require. Therefore, sentiment analysis is first performed on the target nursing plan, and the sentiment semantic features of each segmentation in the target nursing plan are reflected by the sentiment score value, and the keywords in the target nursing plan are extracted. The characteristic values ​​are described to reflect the sentiment semantic features of each keyword. The sentiment semantic features of each segmentation in the local range are then used to reflect the differences in the content information of the nursing plans between historical patients through the difference in nursing information, and then the historical patients with similar nursing plans are divided into the same cluster, eliminating the interference of different factors such as physique and living habits between patients on the personalized recommendation of subsequent nursing plans. At the same time, the similarity of the symptom manifestations between the patient to be tested and the historical patients in different clusters is taken into account. In the process of recommending nursing plans, the nursing plans of historical patients with similar symptoms to the patient to be tested need to be recommended to the patient to be tested. Therefore, the similarity and correlation of the symptom manifestations between the patient to be tested and the historical patients in the target cluster are reflected through the degree of association, and then the nursing plans of the historical patients in each cluster are personalized and targeted for the patient to be tested according to different priorities, thereby improving the effect of nursing information management. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 A flow chart of a method for intelligent management of neurology nursing information provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a neurological nursing information intelligent management method proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0018] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0019] The specific scheme of the intelligent management method of neurology nursing information provided by the present invention is described in detail below with reference to the accompanying drawings.

[0020] See also Figure 1 , which shows a flow chart of a method for intelligent management of neurology nursing information provided by an embodiment of the present invention, the method comprising: Step S1: Obtain the electronic medical records and different types of nursing plans of each historical patient in the Department of Neurology and the electronic medical records of the patient to be tested, perform word segmentation processing on the electronic medical records and nursing plans of the patient to be tested and each historical patient, and obtain the word segmentation of the electronic medical records and nursing plans.

[0021] The recommendation of neurology nursing plan as one of the ways to manage neurology nursing information is helpful to improve treatment effect, enhance patient satisfaction, optimize resource utilization and promote patient recovery. In the process of nursing neurology patients, medical staff will conduct multi-faceted inquiries and tests on patients, such as condition inquiries, CT or ultrasound imaging tests, etc., so as to integrate the collected information about the patient's physical condition into the patient's electronic medical record, and then formulate various types of nursing plans for patients based on the patient's electronic medical record to promote the patient's recovery, and store the patient's electronic medical record together with the nursing plan in the hospital's terminal database for subsequent use and analysis.

[0022] Therefore, the embodiment of the present invention first obtains the electronic medical records of each historical patient in the neurology department and different types of nursing plans and the electronic medical records of the patients to be tested from the terminal database of the hospital, wherein the different types of nursing plans mainly include, for example, drug care plans, including drug information provided to promote rehabilitation, rehabilitation care plans, including guidance information such as limb function exercises, language rehabilitation training and cognitive rehabilitation training provided to promote rehabilitation, and psychological care plans, including psychological counseling information provided to promote rehabilitation.

[0023] It should be noted that the nursing plan for historical patients is known, while the nursing plan for the patient to be tested is unknown. The nursing plan for historical patients needs to be recommended to the patient to be tested later.

[0024] Since the electronic medical records and nursing plans of historical patients and patients to be tested are usually in text form, the embodiment of the present invention also needs to perform word segmentation processing on the electronic medical records and nursing plans of the patients to be tested and each historical patient respectively, and obtain the word segmentations of the electronic medical records and nursing plans, so as to facilitate the subsequent analysis of the differences in electronic medical records and nursing plans between patients based on the word segmentation. After the word segmentation processing, each electronic medical record has multiple word segmentations, and each nursing plan also has multiple word segmentations.

[0025] Preferably, in one embodiment of the present invention, the existing Jieba word segmentation tool is used to perform word segmentation processing on the electronic medical records and nursing plans of the patient to be tested and each historical patient to obtain the word segmentation of the electronic medical records and nursing plans. In other embodiments of the present invention, other word segmentation methods may also be used for word segmentation processing, which is not limited here.

[0026] Step S2: Take any type of nursing plan for any historical patient as the target nursing plan, perform sentiment analysis on the target nursing plan, and obtain the sentiment score value of each segmentation in the target nursing plan; perform keyword extraction on the target nursing plan to obtain the keywords of the target nursing plan; obtain the descriptive feature value of each keyword of the target nursing plan based on the sentiment score value of each segmentation in the preset neighborhood of each keyword in the target nursing plan; obtain the difference in nursing information between any two historical patients based on the difference in descriptive feature values ​​of the same keywords between the same type of nursing plans of any two historical patients and the number of the same keywords; cluster all historical patients based on the difference in nursing information to obtain multiple clustering clusters.

[0027] Due to the differences in physical constitutions and living habits among different patients, that is, there are individual differences among patients. Even patients with similar conditions or symptoms may require certain differences in the nursing plans they require. If the nursing plans of historical patients with similar electronic medical records to the patient to be tested are directly recommended to the patient to be tested, the final recommendation effect will be reduced. Therefore, the embodiment of the present invention eliminates the interference of individual differences by dividing historical patients with similar nursing plans into the same cluster, thereby making targeted and personalized recommendations for the nursing plans of the patient to be tested, thereby improving the recommendation effect and the effect of nursing information management.

[0028] Before clustering historical patients, it is necessary to analyze the information differences of nursing plans between historical patients. The embodiment of the present invention first takes any type of nursing plan for any historical patient as the target nursing plan. Since the nursing plan is provided by medical staff, the content information in the nursing plan includes degree words of emotional expression of medical staff. The difference in emotional semantic features of the content between nursing plans can reflect the information differences between nursing plans. Therefore, sentiment analysis can be performed on the target nursing plan to obtain the sentiment score value of each participle in the target nursing plan. The obtained sentiment score value reflects the sentiment semantic features of each participle in the target nursing plan. Subsequently, based on the difference in the sentiment score values ​​of each participle between the nursing plans, the information differences between the nursing plans are accurately analyzed.

[0029] Preferably, in one embodiment of the present invention, the target care plan is processed using the NTUSD sentiment dictionary to obtain the sentiment score value of each word segment in the target care plan, wherein the NTUSD sentiment dictionary is a technical means well known to those skilled in the art and will not be described in detail here.

[0030] Then, the embodiment of the present invention performs keyword extraction on the target nursing plan to obtain the keywords of the target nursing plan, and obtains the descriptive feature value of each keyword of the target nursing plan according to the sentiment score value of each segmentation in the preset neighborhood of each keyword in the target nursing plan. The descriptive feature value reflects the overall level of the sentiment semantic characteristics of each segmentation in the local range of each key time in the target nursing plan, which is convenient for the subsequent difference in sentiment score values ​​of the same keyword between the same type of nursing plans for each historical patient, accurately analyzes the differences in nursing plan information between each historical patient, and improves the clustering effect. The preset neighborhood of a keyword in the target nursing plan includes the keyword and a preset number of segmentations closest to the keyword, and the preset number is set to 10. That is, the preset neighborhood includes 11 segmentations in total. The specific value of the preset number can also be set by the implementer according to the specific implementation scenario, which is not limited here.

[0031] Preferably, in one embodiment of the present invention, a knowledge vocabulary can be used to extract professional vocabulary from the target care plan as keywords of the target care plan. For example, if the target care plan is a drug care plan, a drug knowledge vocabulary can be used to extract various professional vocabulary about drugs from the target care plan as keywords of the target care plan. Among them, the method of extracting keywords using a knowledge vocabulary is a technical means well known to technical personnel in this field and will not be elaborated here.

[0032] Preferably, in one embodiment of the present invention, the method for obtaining the descriptive feature value of each keyword of the target nursing plan specifically includes: The average of the sentiment score values ​​of all the segmented words in the preset neighborhood of each keyword in the target nursing plan is used as the descriptive feature value of each keyword in the target nursing plan.

[0033] The descriptive feature value of each keyword of each type of nursing plan for each historical patient can be obtained by the same method as above.

[0034] Since each historical patient has multiple types of nursing plans, in the process of analyzing the information differences in nursing plans between historical patients, it is necessary to analyze the nursing plans of the same type between historical patients. The greater the difference in sentiment score values ​​of the same keywords between the same type of nursing plans of historical patients, and the fewer the number of the same keywords between the same type of nursing plans of historical patients, the greater the information difference in nursing plans between historical patients. Therefore, the difference in sentiment score values ​​of the same keywords and the number of the same keywords between the same type of nursing plans of any two historical patients can be analyzed. The difference in content information of nursing plans between any two historical patients can be reflected by the obtained nursing information difference. Subsequently, clustering operations on historical patients can be achieved based on the nursing information difference.

[0035] Preferably, in one embodiment of the present invention, the method for obtaining the difference in nursing information between any two historical patients specifically includes: First, any type of nursing plan is taken as the target type, and the set of all keywords of the nursing plan of the target type of each historical patient is taken as the keyword set of each historical patient about the target type; Then, two patients are arbitrarily selected as the first historical patient and the second historical patient, and the Jaccard correlation coefficient between the keyword set of the first historical patient about the target type and the keyword set of the second historical patient about the target type is negatively mapped to obtain the first information difference coefficient about the target type between the first historical patient and the second historical patient. The larger the first information difference coefficient, the greater the difference in content information of the nursing plan of the target type between the first historical patient and the second historical patient.

[0036] The average of the absolute values ​​of the differences in the descriptive feature values ​​of all the same keywords between the nursing plan of the target type of the first historical patient and the nursing plan of the target type of the second historical patient is taken as the second information difference coefficient regarding the target type between the first historical patient and the second historical patient. The larger the second information difference coefficient is, the greater the difference in the emotional semantic features of the keywords of the nursing plan of the target type between the first historical patient and the second historical patient is, which further indicates that the content information difference of the nursing plan of the target type between the first historical patient and the second historical patient is greater.

[0037] Furthermore, the first information difference coefficient and the second information difference coefficient are integrated to obtain the comprehensive information difference coefficient on the target type between the first historical patient and the second historical patient. The larger the comprehensive information difference coefficient, the greater the information difference in the nursing plan of the target type between the first historical patient and the second historical patient.

[0038] In an embodiment of the present invention, the sum or product of the first information difference coefficient and the second information difference coefficient can be used as the comprehensive information difference coefficient between the first historical patient and the second historical patient regarding the target type to achieve the integration of the two, which is not limited here.

[0039] The same method as above can be used to obtain the comprehensive information difference coefficient of each type between the first historical patient and the second historical patient, and the average value of the comprehensive information difference coefficient of all types between the first historical patient and the second historical patient is used as the nursing information difference between the first historical patient and the second historical patient.

[0040] As an example, in one embodiment of the present invention, the expression of the difference degree of nursing information between the first historical patient and the second historical patient may be specifically, for example, as follows: in, Indicates the difference degree of nursing information between the first historical patient and the second historical patient; Indicates the relationship between the first historical patient and the second historical patient The first information difference coefficient of each type; Indicates the relationship between the first historical patient and the second historical patient The second information difference coefficient of each type; Indicates the relationship between the first historical patient and the second historical patient The comprehensive information difference coefficient of each type; Indicates the number of types of care options; represents the first information difference coefficient between the first historical patient and the second historical patient regarding the target type; represents the Jaccard correlation coefficient between the keyword set of the first historical patient on the target type and the keyword set of the second historical patient on the target type; represents a second information difference coefficient between the first historical patient and the second historical patient regarding the target type; The first target type in the care plan for the first historical patient The descriptive feature value of the keywords; The first in the care plan representing the target type of the second history patient The descriptive feature value of the keyword, among which the first historical patient's target type of nursing plan Keywords and second history of the patient's goal type of care plan The keywords are the same; a number of all common keywords between a care regimen representing a target type for a first historical patient and a care regimen representing a target type for a second historical patient; Indicates the preset adjustment parameter, which is used to prevent the denominator from being 0. The value range is In one embodiment of the present invention, Set to 0.01, The specific value of can also be set by the implementer according to the specific implementation scenario and is not limited here.

[0041] The same method as above can be used to obtain the difference in nursing information between any two historical patients. Then, based on the difference in nursing information, all historical patients can be clustered to obtain multiple clusters, so that historical patients with similar nursing plans can be divided into the same cluster, eliminating the interference of individual differences between patients on the personalized recommendation of subsequent nursing plans.

[0042] Preferably, in one embodiment of the present invention, the method for obtaining multiple clusters specifically includes: The difference in nursing information between any two historical patients is used as the distance measure between any two historical patients. The K-means clustering algorithm is used to cluster all historical patients based on the distance measure between any two historical patients to obtain multiple clusters, where the number of clusters can be determined by the existing elbow method or silhouette coefficient method.

[0043] So far, multiple clusters have been obtained.

[0044] Step S3: Take any cluster as the target cluster, and obtain the degree of association between the patient to be tested and the target cluster according to the number of times each word segment of the electronic medical record of the patient to be tested appears in the electronic medical record of each historical patient in the target cluster, and the number of identical word segments in the electronic medical records of the patient to be tested and each historical patient in the target cluster.

[0045] Since the similarity of the symptoms between the patient to be tested and the historical patients in different clusters is different, in the process of recommending the nursing plan, it is necessary to recommend the nursing plan of the historical patients with similar symptoms to the patient to be tested to the patient to be tested. Therefore, the embodiment of the present invention first analyzes any cluster, and takes any cluster as the target cluster. The more times each segmentation of the electronic medical record of the patient to be tested appears in the electronic medical record of each historical patient in the target cluster, and the more the number of the same segmentations in the electronic medical records between the patient to be tested and each historical patient in the target cluster is, it indicates that the electronic medical record of the patient to be tested and each historical patient in the target cluster are similar. The more similar the content information of the electronic medical records is, the more similar the symptom manifestations between the test patient and the historical patients in the target cluster are. Therefore, the number of times each segmentation of the electronic medical record of the test patient appears in the electronic medical record of each historical patient in the target cluster, as well as the number of identical segmentations in the electronic medical records of the test patient and each historical patient in the target cluster can be analyzed. The obtained degree of correlation can reflect the similarity and correlation of the symptom manifestations between the test patient and the historical patients in the target cluster. Subsequently, based on the degree of correlation, targeted and personalized recommendations can be made for the nursing plan for the test patient, thereby improving the effect of nursing information management in the Department of Neurology.

[0046] Preferably, in one embodiment of the present invention, the method for obtaining the degree of association between the patient to be tested and the target cluster specifically includes: First, any segmentation in the electronic medical record of the patient to be tested is taken as the target segmentation, and the reference historical patients of the target segmentation are selected from all the historical patients in the target cluster. The target segmentation appears at least once in the electronic medical record of the reference historical patient. The more historical patients the target segmentation refers to and the more times the target segmentation appears in the electronic medical record of each reference historical patient, the more commonly the target segmentation is distributed in the electronic medical records of each historical patient in the target cluster. The comprehensive distribution breadth of the target segmentation with respect to the target cluster is obtained according to the number of historical patients referenced by the target segmentation and the number of times the target segmentation appears in the electronic medical record of each reference historical patient. The same method as above can be used to obtain the comprehensive distribution breadth of each segmentation in the electronic medical record of the patient to be tested with respect to the target cluster, and then the average value of the comprehensive distribution breadth of all segmentations in the electronic medical record of the patient to be tested with respect to the target cluster is taken as the overall distribution breadth of the patient to be tested with respect to the target cluster. The larger the overall distribution breadth, the more commonly and widely each segmentation in the electronic medical record of the patient to be tested is distributed in the electronic medical records of the historical patients in the target cluster, and then the more similar the symptom manifestations between the patient to be tested and the historical patients in the target cluster are.

[0047] Preferably, in one embodiment of the present invention, the method for obtaining the comprehensive distribution breadth of the target word segmentation with respect to the target clustering cluster specifically includes: The number of all reference historical patients of the target participle is taken as the numerator, the number of all historical patients in the target cluster is taken as the denominator, the ratio is taken as the first distribution breadth of the target participle in the target cluster, the average number of times the target participle appears in the electronic medical records of all reference historical patients is taken as the second distribution breadth of the target participle in the target cluster, the first distribution breadth and the second distribution breadth are combined to obtain the comprehensive distribution breadth of the target participle with respect to the target cluster.

[0048] In the embodiment of the present invention, the sum or product of the first distribution breadth and the second distribution breadth may be used as the comprehensive distribution breadth of the target word segmentation with respect to the target cluster to achieve integration of the two, which is not limited here.

[0049] As an example, in one embodiment of the present invention, the expression of the comprehensive distribution breadth of the target word segmentation with respect to the target cluster cluster may be specifically, for example: in, Indicates the comprehensive distribution breadth of the target word with respect to the target cluster; The number of all reference history patients representing the target participle; Represents the number of all historical patients in the target cluster; Indicates the first distribution breadth of the target word in the target cluster; Indicates that the target word is in The number of times a reference history patient appears in the electronic medical record; Indicates the second distribution breadth of the target word in the target cluster.

[0050] Then, the set consisting of all the segmentations of the electronic medical records of the patient to be tested is used as the first segmentation set of the patient to be tested, any historical patient in the target cluster is used as the target historical patient, and the set consisting of all the segmentations of the electronic medical records of the target historical patient is used as the second segmentation set of the target historical patient; The Jaccard correlation coefficient between the first segmentation set of the patient to be tested and the second segmentation set of the target historical patient is used as the symptom similarity between the patient to be tested and the target historical patient. The greater the symptom similarity, the more similar the symptom manifestations between the patient to be tested and the target historical patient. The same method as above can be used to obtain the symptom similarity between the patient to be tested and each historical patient in the target cluster. Then, the average value of the symptom similarity between the patient to be tested and all historical patients in the target cluster can be used as the overall similarity between the patient to be tested and the target cluster. The greater the overall similarity, the more similar the symptom manifestations between the patient to be tested and each historical patient in the target cluster.

[0051] Then, the overall similarity between the patient to be tested and the target cluster and the overall distribution breadth of the patient to be tested with respect to the target cluster are integrated and normalized, and the calculation results are limited to Within the range, the degree of association between the patient to be tested and the target cluster is obtained.

[0052] In an embodiment of the present invention, the integration of the two can be achieved by calculating the sum or product of the overall similarity between the patient to be tested and the target cluster and the overall distribution breadth of the patient to be tested with respect to the target cluster, which is not limited here.

[0053] In one embodiment of the present invention, the normalization process may specifically be, for example, maximum and minimum value normalization process. In other embodiments of the present invention, other normalization methods may be selected according to a specific range of values, which will not be described in detail.

[0054] As an example, in one embodiment of the present invention, the expression of the degree of association between the patient to be tested and the target cluster may be specifically, for example, as follows: in, Indicates the degree of association between the patient to be tested and the target cluster; Represents the overall distribution breadth of the patients to be tested with respect to the target cluster; Represents the overall similarity between the patient to be tested and the target cluster; Represents the patient to be tested and the target cluster The similarity of symptoms between historical patients; Represents the number of all historical patients in the target cluster; Represents the normalization function.

[0055] The degree of association between the patient to be tested and each cluster can be obtained by the same method as above.

[0056] Step S4: Based on the degree of association between the patient to be tested and each cluster, the nursing plans of the historical patients in each cluster are recommended to the patient to be tested according to different priorities.

[0057] The greater the degree of correlation between the patient to be tested and each cluster, the greater the similarity and correlation between the symptom manifestations of the patient to be tested and the historical patients in each cluster. In this case, the nursing plans of the historical patients in the clusters with a larger degree of correlation need to be recommended to the patient to be tested first, so as to make targeted and personalized recommendations for the nursing plans of the patient to be tested and improve the effect of nursing information management.

[0058] Preferably, in one embodiment of the present invention, recommending nursing plans of historical patients in each cluster to the patient to be tested according to different priorities includes: According to the order of the correlation between the patient to be tested and each cluster from large to small, the nursing plan of the historical patients in each cluster is recommended to the patient to be tested.

[0059] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0060] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for intelligent management of neurology nursing information, characterized in that: The method comprises: Obtain the electronic medical records and different types of nursing plans of each historical patient in the Department of Neurology and the electronic medical records of the patient to be tested, perform word segmentation processing on the electronic medical records and nursing plans of the patient to be tested and each historical patient, and obtain the word segmentation of the electronic medical records and nursing plans; Taking any type of nursing plan of any historical patient as the target nursing plan, performing sentiment analysis on the target nursing plan, and obtaining the sentiment score value of each segmentation in the target nursing plan; performing keyword extraction on the target nursing plan, and obtaining the keywords of the target nursing plan; obtaining the descriptive feature value of each keyword of the target nursing plan according to the sentiment score value of each segmentation in the preset neighborhood of each keyword in the target nursing plan; obtaining the difference in nursing information between any two historical patients according to the difference in the descriptive feature value of the same keyword between the same type of nursing plans of any two historical patients and the number of the same keywords; clustering all historical patients based on the difference in nursing information to obtain multiple clustering clusters; Taking any cluster as the target cluster, obtaining the degree of association between the patient to be tested and the target cluster according to the number of times each word of the electronic medical record of the patient to be tested appears in the electronic medical record of each historical patient in the target cluster, and the number of identical words in the electronic medical records of the patient to be tested and each historical patient in the target cluster; Based on the degree of association between the patient to be tested and each cluster, the nursing plans of the historical patients in each cluster are recommended to the patient to be tested according to different priorities.

2. A method for intelligent management of neurology nursing information according to claim 1, characterized in that: The descriptive feature value of each keyword of the target nursing plan includes: The average of the sentiment score values ​​of all the segmented words in the preset neighborhood of each keyword in the target nursing plan is used as the descriptive feature value of each keyword in the target nursing plan.

3. The method for intelligent management of neurology nursing information according to claim 1, characterized in that: The obtaining of the difference in nursing information between any two historical patients comprises: Take any type as the target type, and take the set of all keywords of the nursing plan of the target type of each historical patient as the keyword set of each historical patient about the target type; Randomly select two patients as a first historical patient and a second historical patient, perform negative correlation mapping on the Jaccard correlation coefficient between the keyword set of the first historical patient on the target type and the keyword set of the second historical patient on the target type, and obtain a first information difference coefficient between the first historical patient and the second historical patient on the target type; The average of the absolute values ​​of the differences of the descriptive feature values ​​of all the same keywords between the nursing plan of the target type of the first historical patient and the nursing plan of the target type of the second historical patient is taken as the second information difference coefficient about the target type between the first historical patient and the second historical patient; The first information difference coefficient and the second information difference coefficient are integrated to obtain a comprehensive information difference coefficient about the target type between the first historical patient and the second historical patient; The average value of the comprehensive information difference coefficients of all types between the first historical patient and the second historical patient is taken as the nursing information difference degree between the first historical patient and the second historical patient.

4. The method for intelligent management of neurology nursing information according to claim 1, characterized in that: The obtaining of multiple clusters comprises: The difference in nursing information between any two historical patients is used as the distance metric between any two historical patients, and a K-means clustering algorithm is used to cluster all historical patients based on the distance metric between any two historical patients to obtain multiple clusters.

5. The method for intelligent management of neurology nursing information according to claim 1, characterized in that: The step of obtaining the degree of association between the patient to be tested and the target cluster comprises: Taking any participle of the electronic medical record of the patient to be tested as the target participle, selecting a reference historical patient of the target participle from all historical patients in the target cluster, wherein the target participle appears at least once in the electronic medical record of the reference historical patient; According to the number of reference historical patients of the target segmentation word and the number of times the target segmentation word appears in the electronic medical record of each reference historical patient, the comprehensive distribution breadth of the target segmentation word with respect to the target cluster is obtained; the average value of the comprehensive distribution breadth of all segmentations of the electronic medical record of the patient to be tested with respect to the target cluster is taken as the overall distribution breadth of the patient to be tested with respect to the target cluster; The set consisting of all the segmentations of the electronic medical records of the patient to be tested is used as the first segmentation set of the patient to be tested, any historical patient in the target cluster is used as the target historical patient, and the set consisting of all the segmentations of the electronic medical records of the target historical patient is used as the second segmentation set of the target historical patient; The Jaccard correlation coefficient between the first segmentation set of the patient to be tested and the second segmentation set of the target historical patient is used as the symptom similarity between the patient to be tested and the target historical patient; the average value of the symptom similarities between the patient to be tested and all historical patients in the target cluster is used as the overall similarity between the patient to be tested and the target cluster; The overall similarity between the patient to be tested and the target cluster and the overall distribution breadth of the patient to be tested with respect to the target cluster are integrated and normalized to obtain the degree of association between the patient to be tested and the target cluster.

6. A method for intelligent management of neurology nursing information according to claim 5, characterized in that: The step of obtaining the comprehensive distribution breadth of the target word segmentation with respect to the target cluster includes: The number of all reference historical patients of the target participle is used as the numerator, the number of all historical patients in the target cluster is used as the denominator, and the ratio is used as the first distribution breadth of the target participle in the target cluster; The average number of times the target word appears in the electronic medical records of all reference historical patients is taken as the second distribution breadth of the target word in the target cluster; The first distribution breadth and the second distribution breadth are integrated to obtain a comprehensive distribution breadth of the target word segment with respect to the target cluster.

7. The method for intelligent management of neurology nursing information according to claim 1, characterized in that: The method of recommending nursing plans of historical patients in each cluster to the patient to be tested according to different priorities includes: According to the order of the degree of association between the patient to be tested and each cluster from large to small, the nursing plans of the historical patients in each cluster are recommended to the patient to be tested.

8. The method for intelligent management of neurology nursing information according to claim 1, characterized in that: The wording for obtaining electronic medical records and nursing plans includes: Jieba word segmentation tool is used to perform word segmentation on the electronic medical records and nursing plans of the patients to be tested and historical patients to obtain the word segmentation of the electronic medical records and nursing plans.

9. The method for intelligent management of neurology nursing information according to claim 1, characterized in that: The step of obtaining the sentiment score value of each word in the target nursing plan includes: The target nursing plan is processed using the NTUSD sentiment dictionary to obtain the sentiment score value of each word in the target nursing plan.

10. The method for intelligent management of neurology nursing information according to claim 1, characterized in that: The keywords for obtaining the target care plan include: Using the knowledge word bank, professional vocabulary is extracted from the target nursing plan as the keywords of the target nursing plan.

Images