CN116230193B - Intelligent hospital file management method and system - Google Patents
Intelligent hospital file management method and system Download PDFInfo
- Publication number
- CN116230193B CN116230193B CN202310525369.3A CN202310525369A CN116230193B CN 116230193 B CN116230193 B CN 116230193B CN 202310525369 A CN202310525369 A CN 202310525369A CN 116230193 B CN116230193 B CN 116230193B
- Authority
- CN
- China
- Prior art keywords
- case
- index parameter
- value
- index
- degree
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
- G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Public Health (AREA)
- Medical Informatics (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- General Business, Economics & Management (AREA)
- Biomedical Technology (AREA)
- Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention relates to the technical field of data processing, in particular to an intelligent file management method and system for a hospital, comprising the following steps: obtaining basic reference characteristics according to the differences among all index parameters of each case; obtaining extreme distribution characteristics of all index parameters of each case according to standard deviation of each index parameter of each case; obtaining a weight adjustment value of each index parameter of each case according to the extreme distribution characteristics; obtaining local distribution characteristics of each index parameter of each case according to the local aggregation segment of each index parameter of each case; obtaining the value degree of each index parameter of each case according to the weight adjustment value and the local distribution characteristics; obtaining the reference degree of each case according to the basic reference characteristics of each case and the value degree of the index parameters of part of each case; and adjusting the K neighbor distance according to the reference degree of each case. The invention has higher reference value and practical significance of increasing the similarity of cases.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to an intelligent file management method and system for a hospital.
Background
With the development of society, the demand of people for hospitals is higher and higher, so that more and more patient data are stored in a hospital system. And a large amount of similar case data is convenient for doctors, so that the doctors can provide a large amount of referenceable cases when diagnosing the illness state of the patients, and the analysis and the understanding of the current illness state are more reasonable. While similar cases are typically found by K-nearest neighbor algorithm. The hospital archives database is conventionally based on K-means clustering, and according to the severity of the corresponding pathology in the stored multiple cases, and corresponding personal information such as height, weight, past medical history and the like, the hospital archives database is used as a distance measurement feature during clustering. For such data of symptoms, index parameters with higher values are not lacking, such as pepsinogen values in gastritis are generally and abnormally higher, or part of index parameters are influenced by special physical factors of individuals, so that clustering results are easily influenced by such extreme data characteristics, the clustering classification is inaccurate, and further, when a K nearest neighbor algorithm is performed, the characteristics for measuring the distance are inaccurate, and the search of similar cases is influenced.
Disclosure of Invention
The invention provides an intelligent file management method and system for a hospital, which are used for solving the existing problems.
The invention discloses an intelligent file management method and system for a hospital, which adopts the following technical scheme:
one embodiment of the present invention provides an intelligent hospital archive management method, which includes the following steps:
acquiring case information of a current case and a database of corresponding symptoms;
obtaining basic reference characteristics according to the differences among index parameters of each case in the database; obtaining the extreme distribution characteristics of each index parameter of each case according to the difference of each index parameter of each case; obtaining a weight adjustment value of each index parameter of each case according to the extreme distribution characteristics;
acquiring a number average value corresponding to each index parameter, and recording the index parameter of which the corresponding number exceeds the number average value as a first index parameter; the index parameters, the corresponding number of which exceeds the number average value and is adjacent to the first index parameter, are marked as second index parameter pairs; carrying out connection judgment on the numerical range of the index parameter contained by each second index parameter to obtain each section of connected index parameter fragment, and marking the connected index parameter fragment as a local aggregation fragment of each case; obtaining local distribution characteristics of each index parameter of each case according to the local aggregation segments of each case; obtaining the value degree of each index parameter of each case according to the weight adjustment value and the local distribution characteristics;
obtaining the reference degree of each case and the current case when judging the similarity degree according to the basic reference characteristics of each case and the value degree of the index parameters of part of each case;
and adjusting the K neighbor distance according to the reference degree of each case and the current case when judging the similarity degree, and searching the similar cases of the adjusted cases.
Further, the obtaining expression for obtaining the basic reference feature according to the difference between each index parameter of each case in the database is as follows:
wherein the method comprises the steps ofA j-th index parameter representing a corresponding i-th condition in all cases except the current case;a j index parameter representing the current case; j represents J index parameters of the current case;represents the underlying reference feature of the ith condition.
Further, the method for obtaining the extreme distribution characteristics of each index parameter of each case is as follows:
the standard deviation of each index parameter is obtained and marked as a first standard deviation, the maximum value of the standard deviation of all the index parameters is obtained and marked as a maximum standard deviation, and the minimum value of the standard deviation of all the index parameters is obtained and marked as a minimum standard deviation; the result of subtracting the first standard deviation from the maximum standard deviation is recorded as a first difference value; marking the result of subtracting the minimum standard deviation from the first standard deviation as a second difference; comparing the first difference value with the second difference value, and selecting the minimum value as the extreme distribution characteristic.
Further, the method for obtaining the weight adjustment value of each index parameter of each case is as follows:
obtaining the maximum value of the standard deviation of all index parameters, marking the maximum standard deviation, obtaining the minimum value of the standard deviation of all index parameters, marking the minimum standard deviation; the difference value obtained by subtracting the minimum standard deviation from the maximum standard deviation is recorded as a third difference value; and recording the calculation result of the ratio of the extreme distribution characteristic to the third difference value as a weight adjustment value of each index parameter.
Further, the expression for obtaining the local distribution characteristics of each index parameter in each case is as follows:
wherein N represents the total number of locally aggregated fragments;a central value of index parameters corresponding to the nth local aggregation segment is represented;a mean value representing the central values of the plurality of corresponding index parameters except the nth locally aggregated segment;represent the firstThe total number of index parameter values of the local aggregation segments corresponds to the total number;a mean value representing the number of index parameters for all locally aggregated segments;representing the local distribution characteristics obtained in the j index parameter;representing a linear normalization.
Further, the method for obtaining the value degree of each index parameter of each case is as follows:
carrying out linear normalization on the local distribution characteristics of each index parameter, and marking the calculation result as a first characteristic; and recording the product result of the first characteristic and the weight adjustment value as the value degree of each index parameter.
Further, the expression for obtaining the reference degree of each case and the current case when judging the similarity degree is as follows:
wherein the method comprises the steps ofA j-th index parameter representing a corresponding i-th condition in all cases except the current case;a j index parameter representing the current case; j represents J index parameters of the current case;indicating the value degree of the j index parameter;and the reference degree of similarity between the ith case after improvement and the current case is represented.
Further, the specific process of adjusting the K-nearest neighbor distance according to the reference degree of each case and the current case when judging the similarity degree is as follows:
acquiring the original K neighbor distance of each case and marking the original K neighbor distance as an original distance; obtaining the reference degree of each case, carrying out linear normalization processing on the reference degree, and recording the processed reference degree as a first reference degree; recording the difference value obtained by subtracting the first reference degree from 1 as an adjustment coefficient; and (5) marking the product result of the original distance and the adjustment coefficient as the K neighbor distance after adjustment of each case.
Another embodiment of the present invention provides an intelligent hospital archive management system, which includes a case data acquisition module, a case value degree acquisition module, a case reference degree acquisition module, and a K-nearest neighbor distance adjustment module, wherein:
the case data acquisition module acquires case information of the current case and a database of corresponding symptoms;
the case value degree acquisition module is used for acquiring basic reference characteristics according to the difference between each index parameter of each case in the database; obtaining the extreme distribution characteristics of each index parameter of each case according to the difference of each index parameter of each case; obtaining a weight adjustment value of each index parameter of each case according to the extreme distribution characteristics;
acquiring a number average value corresponding to each index parameter, and recording the index parameter of which the corresponding number exceeds the number average value as a first index parameter; the index parameters, the corresponding number of which exceeds the number average value and is adjacent to the first index parameter, are marked as second index parameter pairs; carrying out connection judgment on the numerical range of the index parameter contained by each second index parameter to obtain each section of connected index parameter fragment, and marking the connected index parameter fragment as a local aggregation fragment of each case; obtaining local distribution characteristics of each index parameter of each case according to the local aggregation segments of each case; obtaining the value degree of each index parameter of each case according to the weight adjustment value and the local distribution characteristics;
the case reference degree acquisition module is used for acquiring the reference degree of each case and the current case when judging the similarity degree according to the basic reference characteristics of each case and the value degree of the index parameters of part of each case;
and the K neighbor distance adjusting module is used for adjusting the K neighbor distance according to the reference degree of each case and the current case when judging the similarity degree, and searching the similar cases of the adjusted cases.
The technical scheme of the invention has the beneficial effects that: according to the method, in a plurality of index parameters of cases in a database, each index parameter is analyzed according to the overall distribution characteristics and the local distribution characteristics of the data of the index parameters, the reference value existing in different index parameters is analyzed, the difference of each index parameter among the cases is combined as the basic similarity degree, the reference degree of each case and the current case is obtained finally, the distance measurement characteristics during clustering are adjusted, so that the case finally obtained according to K neighbor calculation has higher reference value, and the practical significance of the case similarity degree is increased.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart showing steps of an intelligent hospital file management method according to the present invention;
FIG. 2 is a block diagram of an intelligent hospital file management system according to the present invention;
FIG. 3 is a diagram showing the relationship between the number and the value corresponding to an index parameter according to the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the present invention to achieve the preset purposes, the following detailed description refers to specific embodiments, structures, features and effects of an intelligent hospital file management method and system according to the present invention with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
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.
The following specifically describes a specific scheme of the intelligent file management method for hospitals provided by the invention with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of steps of an intelligent file management method for a hospital according to an embodiment of the present invention is shown, the method includes the following steps:
step S001: a database of case information and corresponding disorders is obtained.
With the development of society, the demand of people for hospitals is higher and higher, so that more and more patient data are stored in a hospital system. And a large amount of similar case data is convenient for doctors, so that the doctors can provide a large amount of referenceable cases when diagnosing the illness state of the patients, and the analysis and the understanding of the current illness state are more reasonable. While similar cases are typically found by K-nearest neighbor algorithm. The hospital archives database is conventionally based on K-means clustering, and according to the severity of the corresponding pathology in the stored multiple cases, and corresponding personal information such as height, weight, past medical history and the like, the hospital archives database is used as a distance measurement feature during clustering. For such data of symptoms, index parameters with higher values are not lacking, such as pepsinogen values in gastritis are generally and abnormally higher, or part of index parameters are influenced by special physical factors of individuals, so that clustering results are easily influenced by such extreme data characteristics, the clustering classification is inaccurate, and further, when a K nearest neighbor algorithm is performed, the characteristics for measuring the distance are inaccurate, and the search of similar cases is influenced.
The present embodiment does not address a certain disorder, but is described with respect to a disorder such as a stomach disease. Firstly, selecting a patient suffering from stomach diseases, collecting a blood sample of the patient, placing the blood sample into a blood routine analyzer for automatic detection to obtain a plurality of detection index parameters of the current case, preliminarily determining a database corresponding to stomach diseases according to the current case to obtain a plurality of detection index parameters corresponding to all cases, and simultaneously obtaining historical illness information and personal basic information of the patient suffering from the stomach diseases, for example: gender, age, etc.; and classifying the clusters of all cases in a database according to the K-means clusters to obtain a plurality of clusters, and obtaining all cases in the clusters to which the current cases belong.
And then obtaining basic similar characteristics according to the differences among the index parameters in each case, and adjusting contribution degrees in the basic similar characteristics according to the commonality problems corresponding to each index parameter in the database to obtain the reference degree of each case. The method comprises the steps of taking case information of a current case as a clustering center, combining basic information such as age, sex, height, physical sign and the like of case basic information of a patient and information such as past medical history, family medical history and the like of medical history to perform K nearest neighbor calculation, and adjusting a K nearest neighbor distance L obtained by calculating each case and the current case, so that an obtained K nearest neighbor distance result accords with a real reference value.
So far, K-means clustering is carried out on all cases of a database in a hospital system to obtain a plurality of clusters; acquiring all cases in a cluster to which a current case belongs; and calculating K neighbor distances between each case and the current case through case basic feature information, and adjusting the neighbor distances which are not proper originally, so that the obtained similar cases with the current case are more similar.
It should be noted that the problem of commonality of index parameters of the same disorder refers to that index parameters of one or more of the disorder are generally higher or lower, and blood glucose indexes corresponding to all diabetics are generally higher; the corresponding white blood cell count is generally higher for all patients with gastritis.
Step S002: and obtaining basic similar characteristics according to the differences among the index parameters in each case, and obtaining the value degree of the index parameters of each case according to the commonality problem corresponding to each index parameter in the database.
It should be noted that in conventional hospital archive management, the case archives stored in the database are generally classified by a K-means clustering algorithm based on case information, and when a doctor inputs current patient case information in the system, the system outputs similar case characteristics through a K-nearest neighbor algorithm in the corresponding clustering classification. The system passes through the K neighbor algorithm in the corresponding cluster classification: and preliminarily determining the severity of the current symptoms, the difference among the detected multiple index parameters, and the similarity among personal basic information in the aspects of age, sign and the like as distance measurement features, so as to perform K neighbor distance measurement. However, the parameter indexes in a plurality of diseases are greatly affected by the common problems existing in the parameter indexes of the corresponding diseases and the personal factors with serious parts, so that the reference value of the reference indexes in the corresponding diseases is greatly reduced, clustering of the corresponding diseases is caused, and the poor clustering result effect is generated due to the influence of the extreme values or the values with serious influence of the personal factors, so that the distance calculation result of the index parameters is inaccurate when the K neighbor algorithm is performed in the current case, and the weight of the parameters in the distance measurement feature when the K neighbor algorithm is performed is required to be adjusted.
It should be further noted that, in the present embodiment, among a plurality of cases stored in the present hospital system, a case most similar to the case condition of the present doctor is found, and the case most similar to the case condition is referred to, so as to perform more accurate diagnosis on the present case and prediction of the future treatment result. To achieve the above-described expected effects, it is necessary to obtain preliminary similar features between the current case and each case based on the difference between the detection index in each case and the detection index corresponding to the current case in the cluster to which the current case belongs.
Firstly, J index parameters of the corresponding ith case in all cases except the current case are obtained, and basic reference characteristics are calculated according to the difference between the J index parameters of the ith case and the corresponding J index parameters of the current case symptoms:
wherein the method comprises the steps ofA j-th index parameter representing a corresponding i-th condition in all cases except the current case;a j index parameter representing the current case; j represents J index parameters of the current case;the basic reference feature representing the ith condition, the smaller the basic reference feature, the smaller the difference between the ith condition and the current condition, the higher the degree to which the ith condition is used by the doctor as a reference case for the current case.
Because the symptoms of different cases are generated by the comprehensive influence of various factors in the body of a patient, the detection of multiple index parameters of the body is needed, and the detection results of the multiple related index parameters can reflect the severity of the development of the symptoms of the cases and the characteristics of parallel appearance and the like most intuitively. When the multiple related indexes detected by any two cases are almost identical, no matter whether other cases have special factors or not, the similarity degree between the two cases is the closest, so that the accumulated result of the size difference of the specific index parameters can represent the difference between the two cases, when the value is smaller, the difference between the two cases is smaller, the degree of the difference between the two cases serving as a reference case of the current case is higher for a doctor, and the contribution degree of the K nearest neighbor distance calculation is higher.
So far, the basic reference characteristics of all cases in the cluster to which the current case belongs can be obtained by the method。
Furthermore, in the existing cases, based on the detection index parameters under the same condition, certain index parameters are very high, and the index parameters have a larger common problem, so that the index parameters have lower reference value when similar characteristic measurement is carried out; among the detected index parameters, certain index parameter distribution is completely discrete, and the numerical value of the index parameters is greatly influenced by the aspects of personal physical signs and the like, so that the index parameters have lower reference value. However, for the part that exists, namely the aggregation characteristic that the past medical history causes the index parameters to generate the offset in different directions in the local aggregation segment, the aggregation characteristic is different in meaning because of the offset in different directions, so that the setting of different weights is further required for each index parameter in the detected index parameters based on the characteristic.
The weight adjustment value obtained by the integral distribution characteristic of the corresponding j index parameter in all cases except the current case is obtained, and the specific process is as follows:
firstly, according to the current case, a doctor initially diagnoses a classification database corresponding to the symptoms, for example, the current patient initially diagnoses gastritis, and the selected classification database is the database corresponding to the stomach symptoms. And extracting the data of the corresponding j index parameters in all cases except the current case from the current database. The distribution characteristics existing in the j index parameters are calculated by standard deviationThe distribution dispersion condition of the jth index parameters in the current database can be obtained, when the standard deviation of the jth index parameters is larger, the distribution of the jth index parameters in the corresponding database is more dispersed, and the degree of personal influence in the jth index parameters is larger, so that the jth index parameters have lower reference value. When the standard deviation of the jth index parameters is smaller, the jth index parameters are more concentrated in distribution in the corresponding database, the jth index parameters have extremely high commonality problem in the current symptoms, and the reference value is not high.
According to the data of the corresponding j index parameters in all cases except the current case, calculating the extreme distribution characteristics of standard deviations of the j index parameters, namely:
wherein the method comprises the steps ofRepresenting standard deviations of corresponding j index parameters in all cases except the current case;maximum value of standard deviation of index parameters representing all J terms;minimum value of standard deviation of index parameters representing all J terms;and the extreme distribution characteristics of the corresponding j index parameters in all cases except the current case are represented.
The extreme distribution characteristics of standard deviations of all index parameters can be obtained through the method.
It should be added that if there isIf the corresponding j index parameter is the same, then arbitrarily selecting one value as the extreme distribution characteristic of the corresponding j index parameter in all cases except the current case。
In order to represent extreme characteristics of the corresponding jth index parameter and the rest index parameters in the discrete distribution characteristics in all cases except the current case, therefore, the ratio of the extreme distribution characteristics of the jth index parameter to the standard deviation maximum and minimum difference values of all index parameters is selected as a weight adjustment value obtained according to the data distribution characteristics, and the weight adjustment value obtained according to the standard deviation obtained according to the distribution characteristics of the corresponding jth index parameter in all cases except the current case is obtained, namely:
wherein the method comprises the steps ofRepresenting extreme distribution characteristics of the corresponding j index parameters in all cases except the current case;maximum value of standard deviation of index parameters representing all J terms;minimum value of standard deviation of index parameters representing all J terms;the weight adjustment values corresponding to the jth index parameter in all cases except the current case are shown, and if the weight adjustment values are smaller, the distribution characteristics of the jth index parameter are shown to be extreme: the corresponding j index parameter has lower reference value, either in the form of aggregate distribution or complete discrete.
The weight adjustment values of all index parameters can be obtained through the method.
2. The local distribution characteristics of the corresponding j index parameters in all cases except the current case are obtained, and the specific process is as follows:
in addition, it should be noted that the same condition is divided into conditions of different development conditions such as early stage, medium stage and later stage, and different conditions are reflected on index parameters of all J items, so that a plurality of local aggregation fragments are generated, but due to the influence of other personal factors such as past medical history and the like existing in the cases, the values of certain index parameters are partially offset in different directions, so that a plurality of local aggregation characteristics exist in all J index parameters, the cases corresponding to the local aggregation characteristics are considered to have cases with similar medical histories, and the reference degree of the index parameters is higher.
The number corresponding to all J index parameters of the current case is taken as the abscissa, and the number of times of occurrence of all J index parameter values in all cases, namely the number of points corresponding to each index parameter value, is taken as the ordinate, so that a diagram of the relationship between the number corresponding to the index parameter and the number can be obtained, as shown in fig. 3.
The item numbers of all item index parameters are recorded as a first item number A; counting the number corresponding to the numerical value of each index parameter to obtain the total number corresponding to all index parameters, and marking the total number as a first number B; the ratio of the first number to the first itemAs a first average value, judging whether the number corresponding to each index parameter value exceeds the first average value, and if the number corresponding to each index parameter value exceeds the first average value, judging each index parameter valueAnd if the index parameter is marked as the first index parameter, judging whether the number of the left side and the right side of each first index parameter value corresponding to the index parameter value nearest to the first index parameter is equal to or exceeds a first average value, if so, judging that the first index parameter is communicated, and if not, judging that the first index parameter is not communicated. For example, in fig. 3, o2 is a first index parameter, and o1 and o3 are index parameters nearest to the left and right sides of the first index parameter, where the numbers h1, h2, and h3 corresponding to o1, o2, and o3 are all greater than a first average value h0; the index parameter fragments formed by o1, o2 and o3 are communicated index parameter fragments.
The method can obtain a plurality of segments of communicated index parameter fragments, and takes the index parameter value of each segment of communicated index parameter as a local aggregation fragment.
According to the mode, all N local aggregation segments in all J index parameter values are obtained, and the index parameter center value corresponding to the nth local aggregation segment is calculatedObtaining the distribution characteristics of the local aggregation segments obtained in the j index parameter, namely:
where N represents the total number of locally aggregated segments,a central value of index parameters corresponding to the nth local aggregation segment is represented;a mean value representing the central values of the plurality of corresponding index parameters except the nth locally aggregated segment;indicating the total number corresponding to the index parameter values of the nth partial aggregation segment;the average value of the index parameter quantity of all the local aggregation fragments is represented, the larger the average value is, the more the index parameter quantity corresponding to the local aggregation fragments is, and the higher the contribution degree corresponding to the calculation of the discrete distribution characteristics of the fragments is;representing a linear normalization.The local distribution characteristics obtained in the jth index parameter are represented, the higher the local distribution characteristics are, the larger the deviations of the jth index parameter in different directions of the index parameter caused by factors such as past medical history are, so that the degree of a plurality of discrete local aggregation distribution characteristics formed is more obvious, and when the similarity degree of the jth index parameter among cases is analyzed, the reference value of the jth index parameter among cases is larger when the similarity characteristics exist.
The local aggregation segment distribution characteristics of all index parameters can be obtained through the method.
3. The value degree of the corresponding j index parameter in all cases except the current case is obtained, and the specific process is as follows:
weight adjustment values corresponding to the j index parameters in all cases except the current case obtained according to the steps 1 and 2Local aggregate fragment distribution featuresThe final value degree is obtained for the j index parameters, namely:
wherein the method comprises the steps ofWeight adjustment value representing corresponding j-th index parameter in all cases except current case；Representing local distribution characteristics of the corresponding j index parameters in all cases except the current case;indicating the value degree of the j index parameter; when the overall data distribution characteristic corresponding to the parameter index shows extremely discrete characteristics or is gathered at a local gathering segment, meanwhile, the data in the overall data distribution characteristic does not have the local gathering characteristic, and the corresponding current index parameter has smaller value when carrying out similarity evaluation between cases;representing a linear normalization.
So far, the value degree of all index parameters can be obtained by the method according to the weight adjustment values of all index parameters and the distribution characteristics of the local aggregation segments.
Step S003: and obtaining the reference degree of each case and the current case when judging the similarity degree according to the basic reference characteristics of each case and the value degree of each case.
Then, the reference degree of the i-th case and the current case in judging the similarity degree in all cases except the current case is obtained, namely:
wherein the method comprises the steps ofA j-th index parameter representing the i-th condition in all cases except the current case;a j index parameter representing the current case; j represents J index parameters of the current case;represents item jThe degree of value of the index parameter;and the reference degree of similarity between the ith case after improvement and the current case is represented.
The reference degree of all cases and the current case in judging the similarity degree can be obtained through the method.
It should be noted that, when the difference between each index parameter and the current proportion is regarded as the similar degree, the reference degree corresponding to each index parameter obtained by calculation according to the method is obtained, and the adjustment degree of each case obtained finally according to each index parameter when K neighbor distance measurement is performed with the current case as the center.
Step S004: and carrying out adjustment degree of the K neighbor distance according to the reference degree of each case.
According to the reference degree of each case obtained in step S003, taking the case information of the current case as a clustering center, including basic information such as age, sex, height and the like of the case of the patient and information such as prior medical history, family medical history and the like of the medical history, performing K nearest neighbor calculation, and calculating and adjusting the obtained K nearest neighbor distance between each case and the reference degree of the current case, namely:
wherein the method comprises the steps ofRepresenting the reference degree of the ith case in all cases except the current case;representing the original K nearest neighbor distance between the ith case and the current case in all cases except the current case;representing the adjustment between the ith case and the current case among all cases except the current caseK nearest neighbor distance of (2);representing a linear normalization.
And each case is adjusted to the distance from the current case according to the obtained reference degree. The obtained K neighbor distance result is more in line with the real reference value.
And (3) adjusting the K neighbor distance between all cases except the current disease and the current case, and after improving the similarity between different cases and the current case, searching similar cases of the current case again in a hospital system.
Through the steps, the intelligent file management for the hospital is completed.
Another embodiment of the present invention provides an intelligent hospital archive management system, as shown in fig. 2, comprising the following modules:
the case data acquisition module 101 acquires case information of a current case and a database of corresponding symptoms;
the case value degree obtaining module 102 obtains basic reference characteristics according to the difference between each index parameter of each case in the database; obtaining the extreme distribution characteristics of each index parameter of each case according to the difference of each index parameter of each case; obtaining a weight adjustment value of each index parameter of each case according to the extreme distribution characteristics;
acquiring a number average value corresponding to each index parameter, and recording the index parameter of which the corresponding number exceeds the number average value as a first index parameter; the index parameters, the corresponding number of which exceeds the number average value and is adjacent to the first index parameter, are marked as second index parameter pairs; carrying out connection judgment on the numerical range of the index parameter contained by each second index parameter to obtain each section of connected index parameter fragment, and marking the connected index parameter fragment as a local aggregation fragment of each case; obtaining local distribution characteristics of each index parameter of each case according to the local aggregation segments of each case; obtaining the value degree of each index parameter of each case according to the weight adjustment value and the local distribution characteristics;
the case reference degree obtaining module 103 obtains the reference degree of each case and the current case when judging the similarity degree according to the basic reference feature of each case and the value degree of the index parameter of each case part;
and the K neighbor distance adjustment module 104 is used for adjusting the K neighbor distance according to the reference degree of each case and the current case when judging the similarity degree, and searching the similar cases of the adjusted cases.
According to the embodiment, in a plurality of index parameters of the cases in the database, each index parameter is analyzed according to the overall distribution characteristics and the local distribution characteristics of the data of the index parameter, the reference value existing in different index parameters is analyzed, the difference of each index parameter among the cases is taken as the basic similarity degree, the reference degree of each case and the current case is obtained finally, the distance measurement characteristics during clustering are adjusted, so that the case finally obtained according to K neighbor calculation has higher reference value, and the practical significance of the case similarity degree is increased.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (2)
1. An intelligent file management method for a hospital is characterized by comprising the following steps:
acquiring case information of a current case and a database of corresponding symptoms;
obtaining basic reference characteristics according to the differences among index parameters of each case in the database; obtaining the extreme distribution characteristics of each index parameter of each case according to the difference of each index parameter of each case; obtaining a weight adjustment value of each index parameter of each case according to the extreme distribution characteristics;
acquiring a number average value corresponding to each index parameter, and recording the index parameter of which the corresponding number exceeds the number average value as a first index parameter; the index parameters, the corresponding number of which exceeds the number average value and is adjacent to the first index parameter, are marked as second index parameter pairs; carrying out connection judgment on the numerical range of the index parameter contained by each second index parameter to obtain each section of connected index parameter fragment, and marking the connected index parameter fragment as a local aggregation fragment of each case; obtaining local distribution characteristics of each index parameter of each case according to the local aggregation segments of each case; obtaining the value degree of each index parameter of each case according to the weight adjustment value and the local distribution characteristics;
obtaining the reference degree of each case and the current case when judging the similarity degree according to the basic reference characteristics of each case and the value degree of the index parameters of part of each case;
according to the reference degree of each case and the current case when judging the similarity degree, adjusting the K neighbor distance, and searching the similar cases of the adjusted cases;
the obtained expression of the basic reference feature according to the difference between each index parameter of each case in the database is as follows:
wherein the method comprises the steps ofA j-th index parameter representing a corresponding i-th condition in all cases except the current case; />A j index parameter representing the current case; j represents J index parameters of the current case; />A basal reference signature representing an ith condition;
the method for acquiring the extreme distribution characteristics of each index parameter of each case is as follows:
the standard deviation of each index parameter is obtained and marked as a first standard deviation, the maximum value of the standard deviation of all the index parameters is obtained and marked as a maximum standard deviation, and the minimum value of the standard deviation of all the index parameters is obtained and marked as a minimum standard deviation; the result of subtracting the first standard deviation from the maximum standard deviation is recorded as a first difference value; marking the result of subtracting the minimum standard deviation from the first standard deviation as a second difference; comparing the first difference value with the second difference value, and selecting the minimum value as an extreme distribution characteristic;
the method for obtaining the weight adjustment value of each index parameter of each case is as follows:
obtaining the maximum value of the standard deviation of all index parameters, marking the maximum standard deviation, obtaining the minimum value of the standard deviation of all index parameters, marking the minimum standard deviation; the difference value obtained by subtracting the minimum standard deviation from the maximum standard deviation is recorded as a third difference value; recording the calculation result of the ratio of the extreme distribution characteristic to the third difference value as a weight adjustment value of each index parameter;
the acquisition expression of the local distribution characteristics of each index parameter of each case is as follows:
wherein N represents the total number of locally aggregated fragments;a central value of index parameters corresponding to the nth local aggregation segment is represented; />A mean value representing the central values of the plurality of corresponding index parameters except the nth locally aggregated segment; />Indicating the total number corresponding to the index parameter values of the nth partial aggregation segment; />Index parameters representing all locally aggregated segmentsThe mean value of the number; />Representing the local distribution characteristics obtained in the j index parameter; />Representing a linear normalization;
the method for acquiring the value degree of each index parameter of each case is as follows:
carrying out linear normalization on the local distribution characteristics of each index parameter, and marking the calculation result as a first characteristic; recording the product result of the first characteristic and the weight adjustment value as the value degree of each index parameter;
the obtaining expression of the reference degree of each case and the current case when judging the similarity degree is as follows:
wherein the method comprises the steps ofA j-th index parameter representing a corresponding i-th condition in all cases except the current case; />A j index parameter representing the current case; j represents J index parameters of the current case; />Indicating the value degree of the j index parameter; />A reference degree indicating a degree of similarity of the modified i-th case with the current case;
the specific process of adjusting the K neighbor distance according to the reference degree of each case and the current case when judging the similarity degree is as follows:
acquiring the original K neighbor distance of each case and marking the original K neighbor distance as an original distance; obtaining the reference degree of each case, carrying out linear normalization processing on the reference degree, and recording the processed reference degree as a first reference degree; recording the difference value obtained by subtracting the first reference degree from 1 as an adjustment coefficient; and (5) marking the product result of the original distance and the adjustment coefficient as the K neighbor distance after adjustment of each case.
2. An intelligent hospital file management system is characterized in that the system comprises the following modules:
the case data acquisition module acquires case information of the current case and a database of corresponding symptoms;
the case value degree acquisition module is used for acquiring basic reference characteristics according to the difference between each index parameter of each case in the database; obtaining the extreme distribution characteristics of each index parameter of each case according to the difference of each index parameter of each case; obtaining a weight adjustment value of each index parameter of each case according to the extreme distribution characteristics;
acquiring a number average value corresponding to each index parameter, and recording the index parameter of which the corresponding number exceeds the number average value as a first index parameter; the index parameters, the corresponding number of which exceeds the number average value and is adjacent to the first index parameter, are marked as second index parameter pairs; carrying out connection judgment on the numerical range of the index parameter contained by each second index parameter to obtain each section of connected index parameter fragment, and marking the connected index parameter fragment as a local aggregation fragment of each case; obtaining local distribution characteristics of each index parameter of each case according to the local aggregation segments of each case; obtaining the value degree of each index parameter of each case according to the weight adjustment value and the local distribution characteristics;
the case reference degree acquisition module is used for acquiring the reference degree of each case and the current case when judging the similarity degree according to the basic reference characteristics of each case and the value degree of the index parameters of part of each case;
the K neighbor distance adjustment module is used for adjusting the K neighbor distance according to the reference degree of each case and the current case when judging the similarity degree, and searching the similar cases of the adjusted cases;
the obtained expression of the basic reference feature according to the difference between each index parameter of each case in the database is as follows:
wherein the method comprises the steps ofA j-th index parameter representing a corresponding i-th condition in all cases except the current case; />A j index parameter representing the current case; j represents J index parameters of the current case; />A basal reference signature representing an ith condition;
the method for acquiring the extreme distribution characteristics of each index parameter of each case is as follows:
the standard deviation of each index parameter is obtained and marked as a first standard deviation, the maximum value of the standard deviation of all the index parameters is obtained and marked as a maximum standard deviation, and the minimum value of the standard deviation of all the index parameters is obtained and marked as a minimum standard deviation; the result of subtracting the first standard deviation from the maximum standard deviation is recorded as a first difference value; marking the result of subtracting the minimum standard deviation from the first standard deviation as a second difference; comparing the first difference value with the second difference value, and selecting the minimum value as an extreme distribution characteristic;
the method for obtaining the weight adjustment value of each index parameter of each case is as follows:
obtaining the maximum value of the standard deviation of all index parameters, marking the maximum standard deviation, obtaining the minimum value of the standard deviation of all index parameters, marking the minimum standard deviation; the difference value obtained by subtracting the minimum standard deviation from the maximum standard deviation is recorded as a third difference value; recording the calculation result of the ratio of the extreme distribution characteristic to the third difference value as a weight adjustment value of each index parameter;
the acquisition expression of the local distribution characteristics of each index parameter of each case is as follows:
wherein N represents the total number of locally aggregated fragments;a central value of index parameters corresponding to the nth local aggregation segment is represented; />A mean value representing the central values of the plurality of corresponding index parameters except the nth locally aggregated segment; />Indicating the total number corresponding to the index parameter values of the nth partial aggregation segment; />A mean value representing the number of index parameters for all locally aggregated segments; />Representing the local distribution characteristics obtained in the j index parameter; />Representing a linear normalization;
the method for acquiring the value degree of each index parameter of each case is as follows:
carrying out linear normalization on the local distribution characteristics of each index parameter, and marking the calculation result as a first characteristic; recording the product result of the first characteristic and the weight adjustment value as the value degree of each index parameter;
the obtaining expression of the reference degree of each case and the current case when judging the similarity degree is as follows:
wherein the method comprises the steps ofA j-th index parameter representing a corresponding i-th condition in all cases except the current case; />A j index parameter representing the current case; j represents J index parameters of the current case; />Indicating the value degree of the j index parameter; />A reference degree indicating a degree of similarity of the modified i-th case with the current case;
the specific process of adjusting the K neighbor distance according to the reference degree of each case and the current case when judging the similarity degree is as follows:
acquiring the original K neighbor distance of each case and marking the original K neighbor distance as an original distance; obtaining the reference degree of each case, carrying out linear normalization processing on the reference degree, and recording the processed reference degree as a first reference degree; recording the difference value obtained by subtracting the first reference degree from 1 as an adjustment coefficient; and (5) marking the product result of the original distance and the adjustment coefficient as the K neighbor distance after adjustment of each case.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310525369.3A CN116230193B (en) | 2023-05-11 | 2023-05-11 | Intelligent hospital file management method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310525369.3A CN116230193B (en) | 2023-05-11 | 2023-05-11 | Intelligent hospital file management method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116230193A CN116230193A (en) | 2023-06-06 |
| CN116230193B true CN116230193B (en) | 2023-07-21 |
Family
ID=86589574
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310525369.3A Active CN116230193B (en) | 2023-05-11 | 2023-05-11 | Intelligent hospital file management method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116230193B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20060003741A (en) * | 2004-07-07 | 2006-01-11 | 학교법인 울산공업학원 | Data classification method |
| CN113568044A (en) * | 2020-04-28 | 2021-10-29 | 中国石油天然气股份有限公司 | Method and device for determining arrival time of acoustic array logging head wave |
| CN115858630A (en) * | 2023-02-21 | 2023-03-28 | 新风光电子科技股份有限公司 | Abnormity detection method for energy storage data of energy storage power station |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2141247C1 (en) * | 1998-01-29 | 1999-11-20 | Омский государственный университет | Method for diagnosing cardiac system functional state |
| US9271651B2 (en) * | 2009-11-30 | 2016-03-01 | General Electric Company | System and method for integrated quantifiable detection, diagnosis and monitoring of disease using patient related time trend data |
| CN102495901B (en) * | 2011-12-16 | 2014-10-15 | 山东师范大学 | Method for keeping balance of implementation class data through local mean |
| CN103632154B (en) * | 2013-12-16 | 2018-02-02 | 福建师范大学 | Cicatrix of skin image decision method based on second harmonic analyzing image texture |
| CN113159546B (en) * | 2021-04-12 | 2024-05-14 | 北京工商大学 | Crop supply chain hazard risk judging method and system based on unsupervised dimension reduction density clustering |
| CN114582516A (en) * | 2022-02-25 | 2022-06-03 | 吾征智能技术(北京)有限公司 | Disease multi-source data processing method and device, storage medium and electronic device |
| CN115910362A (en) * | 2022-11-15 | 2023-04-04 | 温州大学 | Atopic dermatitis characteristic prediction method based on enhanced particle swarm optimization |
-
2023
- 2023-05-11 CN CN202310525369.3A patent/CN116230193B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20060003741A (en) * | 2004-07-07 | 2006-01-11 | 학교법인 울산공업학원 | Data classification method |
| CN113568044A (en) * | 2020-04-28 | 2021-10-29 | 中国石油天然气股份有限公司 | Method and device for determining arrival time of acoustic array logging head wave |
| CN115858630A (en) * | 2023-02-21 | 2023-03-28 | 新风光电子科技股份有限公司 | Abnormity detection method for energy storage data of energy storage power station |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116230193A (en) | 2023-06-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109411082B (en) | Medical quality evaluation and treatment recommendation method | |
| CN111261282A (en) | Sepsis early prediction method based on machine learning | |
| CN112201330B (en) | Medical quality monitoring and evaluating method combining DRGs tool and Bayesian model | |
| CN111161814A (en) | DRGs automatic grouping method based on convolutional neural network | |
| CN114639479A (en) | Intelligent diagnosis auxiliary system based on medical knowledge map | |
| CN113838577B (en) | Convenient layered old people MODS early death risk assessment model, device and establishment method | |
| Ma et al. | A new classifier fusion method based on historical and on-line classification reliability for recognizing common CT imaging signs of lung diseases | |
| US20230248998A1 (en) | System and method for predicting diseases in its early phase using artificial intelligence | |
| CN113593708A (en) | Sepsis prognosis prediction method based on integrated learning algorithm | |
| CN111370124A (en) | Health analysis system and method based on facial recognition and big data | |
| CN114023441A (en) | Severe AKI early risk assessment model and device based on interpretable machine learning model and development method thereof | |
| JP7467504B2 (en) | Methods and devices for determining chromosomal aneuploidy and for building classification models - Patents.com | |
| CN117423451B (en) | Intelligent molecular diagnosis method and system based on big data analysis | |
| CN115346598A (en) | Chronic kidney disease genetic gene risk screening system | |
| Balnarsaiah et al. | Parkinson’s disease detection using modified ResNeXt deep learning model from brain MRI images | |
| CN116230193B (en) | Intelligent hospital file management method and system | |
| CN117195027A (en) | Cluster weighted clustering integration method based on member selection | |
| CN113707330B (en) | Construction method of syndrome differentiation model of Mongolian medicine, syndrome differentiation system and method of Mongolian medicine | |
| CN110010246A (en) | A kind of disease Intelligent Diagnosis Technology based on neural network and confidence interval | |
| CN115312196A (en) | Novel model construction evaluation method for screening pressure injury risk factors and application thereof | |
| CN114512240A (en) | Gout prediction model system, equipment and storage medium | |
| Rajmohan et al. | G-Sep: A Deep Learning Algorithm for Detection of Long-Term Sepsis Using Bidirectional Gated Recurrent Unit | |
| CN115132351B (en) | Diagnostic data feedback evaluation system and method based on real world research | |
| Qin | A Prediction Model of Diabetes Based on Ensemble Learning | |
| Hameed | An efficient method of classification the gestational diabetes using id3 classifier |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |