CN109119134A - Medical history data processing method, medical data recommender system, equipment and medium - Google Patents

Medical history data processing method, medical data recommender system, equipment and medium Download PDF

Info

Publication number
CN109119134A
CN109119134A CN201810902948.4A CN201810902948A CN109119134A CN 109119134 A CN109119134 A CN 109119134A CN 201810902948 A CN201810902948 A CN 201810902948A CN 109119134 A CN109119134 A CN 109119134A
Authority
CN
China
Prior art keywords
symptom
information
data
default
relevant
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.)
Pending
Application number
CN201810902948.4A
Other languages
Chinese (zh)
Inventor
范文历
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Maijing (hangzhou) Health Management Co Ltd
Original Assignee
Maijing (hangzhou) Health Management Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Maijing (hangzhou) Health Management Co Ltd filed Critical Maijing (hangzhou) Health Management Co Ltd
Priority to CN201810902948.4A priority Critical patent/CN109119134A/en
Publication of CN109119134A publication Critical patent/CN109119134A/en
Pending legal-status Critical Current

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
    • 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
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • Epidemiology (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Biomedical Technology (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Pathology (AREA)
  • Medical Treatment And Welfare Office Work (AREA)

Abstract

Medical history data processing method, medical data recommender system, equipment and medium of the invention, the medical history data processing method includes: by each case history that a medical history is concentrated and its include that the content of each project is expressed as by mathematical character mode the project information under a structured patient record data and its various classifications of the items for being included, with medical structure medical record data collection of the formation comprising one or more structured patient record data；In turn, medical data recommender system can be carried out based on the medical structure medical record data collection to the data analysis about information such as symptoms, to predict and recommend to meet the interrogation foundation in efficient interrogation direction, Chinese medicinal interrogation process is efficiently assisted, the realistic meaning with height.

Description

Medical history data processing method, medical data recommender system, equipment and medium
Technical field
The present invention relates to intellectual medical technical fields, push away more particularly to medical history data processing method, medical data Recommend system, equipment and medium.
Background technique
Chinese medicinal interrogation is the clinical information acquisition method that communication is directly carried out between doctors and patients, during disease examination Have a very important role, be Chinese medicine hope, hear, asking, one of the important content for cutting the four methods of diagnosis.By interrogation, doctor can just be obtained Comprehensive conditions of patients data is taken, such as the subjective symptoms, both of the generation of disease, development, change procedure, diagnosis and treatment process and patient Toward medical history, personal lifestyle history, family history etc..The process of interrogation, while being also the process of a doctor differentiation thought.In interrogation Cheng Zhong, doctor must be good at carrying out thinking analysis to the conditions of patients information obtained, and be faced according to theory of traditional Chinese medical science and individual Bed experience, tracks new clue, to give farther insight into the state of an illness, accomplishes to debate when asking, ask when debating, ask and debate combination, subtract The blindness of few interrogation, improves the accuracy of diagnosis.
But the management of the doctor's medical history information of centering at present is still based on papery or no data and taxonomically inputs computer etc. In electronic equipment, there is no methods effectively to be managed, to carry out big data analysis.
In addition, it is further, since the conditions of patients data obtained by interrogation has Chinese medical discrimination and medical diagnosis on disease There is important role, therefore the interrogation process needs of " effective ", which can reach, " obtains and realize accurate Chinese medical discrimination and medical diagnosis on disease institute Need conditions of patients data " requirement.In order to conditions of patients information needed for Overall Acquisition, and since state of an illness information is related to model It encloses extensively, successive dynasties famous doctor of Chinese Medicine sums up various interrogation frames to ensure and obtain the comprehensive of interrogation information, such as " ten ask song " Deng.These interrogation frames have certain directive significance, but clinical practice use when, also will according to the specific state of an illness of patient, Flexibly and there is primary and secondary inquire, cannot machine-made machinery find out by asking seemingly casual questions.Simultaneously as doctor can divide in clinical practice The consulting hours of each patient of dispensing are usually very limited, carry out machinery fully according to interrogation frame and find out that there is also use in reality by asking seemingly casual questions Difficulty.
In clinical practice of Chinese medicine, " efficient " interrogation process is realized, and " obtain in finite time and realize accurate Syndrome Differentiation of Chinese Medicine Conditions of patients data needed for card and medical diagnosis on disease " requires the theoretical level, thinking ability and clinical experience of doctor all very high. And the horizontal capability of doctor of traditional Chinese medicine is still different in reality, and common horizontal doctor of traditional Chinese medicine how to be helped to realize efficient interrogation Process is then problem to be solved and the realistic meaning with height.
Summary of the invention
In view of the foregoing deficiencies of prior art, the purpose of the present invention is to provide medical history data processing method, Medical data recommender system, equipment and medium are conducive to analysis for realizing the scientific classification management to medical history data, and It is also able to cooperate big data analysis means auxiliary improvement interrogation efficiency, solves the problems of the prior art.
In order to achieve the above objects and other related objects, the present invention provides a kind of medical history data processing method, comprising: It by each case history that a medical history is concentrated and its include that the content of each project is expressed as one by mathematical character mode Project information under structured patient record data and its various classifications of the items for being included includes one or more described structures to be formed Change the medical structure medical record data collection of medical record data.
In one embodiment of the invention, the classification of the items include: the Symptomatic classification comprising one or more symptom informations, And the relative symptom relevant classification comprising one or more symptom relevant informations；The described method includes: from the medicine structure Change and extract various unduplicated symptom informations in data set, storage forms symptom information library；From the medicine structure data set Various unduplicated symptom relevant informations are extracted in corresponding each symptom relevant classification respectively, and storage forms related to each symptom point The corresponding relevant information library of class.
In one embodiment of the invention, the medicine belongs to traditional Chinese medical science field, and the symptom relevant classification includes: patient's base Any one or more combination in this information, sign, medical history, card type and prescription.
In one embodiment of the invention, the medical history data processing method, comprising: extract in symptom information library Each symptom information between, and/or each symptom information and relevant information library in degree of association data between each symptom relevant information, and add With storage.
Association degree in one embodiment of the invention, between the various symptoms information extracted in symptom information library According to, comprising: every kind of symptom information in corresponding symptom information library establishes symptom and vector occurs；Wherein, the symptom occurs Vector is expressed as the set for the characterization value whether every kind of symptom information appears in each structured patient record data；Calculate various diseases There is the similarity information between vector in symptom corresponding to shape information, as the degree of association data.
In one embodiment of the invention, the calculation method of the similarity information include: COS distance, Euclidean distance, Standardize any one in Euclidean distance, mahalanobis distance, Hamming distance and manhatton distance.
It is each in each symptom information extracted in symptom information library and relevant information library in one embodiment of the invention Degree of association data between symptom relevant information, comprising: count what each symptom relevant information in relevant information library occurred respectively The quantity of the structured patient record data of various symptoms information appearance accounts for structured patient record number respectively under situation and in symptom information library It is the conditional probability of every kind of symptom information occur in the case of every kind of symptom relevant information occurs with approximate representation according to the ratio of collection, As the degree of association data.
In one embodiment of the invention, the medicine belongs to traditional Chinese medical science field, the relevant information library include: comprising one or The card type information bank of multiple card type information, and/or prescription information library comprising one or more prescription informations.
It include main prescription information in the prescription information in one embodiment of the invention；The main prescription information is used for Substitute the prescription information.
In order to achieve the above objects and other related objects, the present invention provides a kind of medical data recommender system, is applied to one It is stored with the processing unit of structured patient record data set；Wherein, the structured patient record data set includes at least one structuring Project information under medical record data, each structured patient record data and its various classifications of the items for being included is to a medicine Case history concentrate a case history and its include each project content mathematical character；The classification of the items includes: comprising one Or Symptomatic classification and the relative symptom relevant classification comprising one or more symptom relevant informations of multiple symptom informations； The processing unit is also stored with: gathering what various unduplicated default symptom informations in the structured patient record data set were formed It is various in default symptom information library and the medicine structure data set that the various classifications of the items are gathered respectively not repeat Default symptom relevant information formed each relevant information library；The processing unit is also stored with: the default symptom information Between each default symptom information in library, and/or in each default symptom information and relevant information library between each default symptom relevant information Degree of association data；Institute's system includes: receiving unit, includes believing to diagnostic symptom for one group of symptom information to be diagnosed for obtaining Cease sequence；Processing unit, for matching and counting in the degree of association data between the default symptom information each in symptom information library Degree of association data of the one group of symptom information to be diagnosed respectively between each default symptom information are calculated, set forms associated diseases Shape sequence, for choosing default symptom information conduct corresponding to wherein one or more maximum data of magnitude of degree of association data Recommend symptom information；And/or the processing unit, for being inputted using the structured patient record data set as data, meter Calculate the posteriority that the default symptom relevant information in each of the relevant information library occurred based on each symptom information to be diagnosed is occurred Probability；Select in the respectively default symptom relevant information posterior probability values maximum one as recommendation symptom relevant information；Or Person judges the maximum multiple default symptom relevant informations of posterior probability values alternately symptom relevant information each alternative Whether the magnitude of the proximity data between symptom relevant information is lower than preset value；If it is not, then selecting posterior probability values maximum standby Select symptom relevant information as recommendation symptom relevant information；If so, in the symptom information library each default symptom information with In degree of association data in relevant information library between each default symptom relevant information, each alternative symptom relevant information is matched respectively and respectively Target association degree evidence between default symptom information, and according to the target association degree being matched to according to calculating: be relevant to it is each Diversity factor data between the alternative symptom relevant information of all each of different default symptom informations of symptom information to be diagnosed, set Formation project distinguishes symptom sequence, pre- corresponding to wherein one or more maximum data of magnitude of diversity factor data for choosing If symptom information is as recommendation symptom information as recommendation symptom information.
In one embodiment of the invention, the degree of association data between the default symptom information include: default symptom information There is the similarity information between vector in symptom corresponding to each default symptom information in library；It is according to right that vector, which occurs, in the symptom Each default symptom information is answered to be established, for indicating whether each default symptom information appears in each structuring The set of characterization value in medical record data；Degree of association data between the default symptom information each in symptom information library In, degree of association data of the one group of symptom information to be diagnosed respectively between each default symptom information are matched and calculated, are collected It closes and forms association symptom sequence, comprising: in the degree of association data between default symptom information, match each described to diagnostic symptom With the current degree of association data between each default symptom information；To each current association relevant to each default symptom information Degree constructs the association symptom sequence according to summed result is respectively obtained, with the set of each summed result；Wherein, in the pass Join in symptom sequence, the magnitude of the summed result is bigger, corresponding to default symptom information as recommendation symptom information Priority is higher.
In one embodiment of the invention, the similarity information is by COS distance, Euclidean distance, standardization Euclidean What any one calculation in distance, mahalanobis distance, Hamming distance and manhatton distance obtained.
It is each in each default symptom information and relevant information library in the symptom information library in one embodiment of the invention Degree of association data between default symptom relevant information, be in the case of each symptom relevant information occurs in relevant information library and The quantity for the structured patient record data that every kind of symptom information occurs in symptom information library accounts for the ratio of structured patient record data set respectively Rate is the conditional probability of every kind of symptom information occur in the case of every kind of symptom relevant information occurs with approximate representation, as described Degree of association data.
In one embodiment of the invention, the multiple alternative symptom relevant information includes: that posterior probability values are maximum and secondary Default symptom relevant information；It is described to be relevant to all each of different from each symptom information to be diagnosed default symptom informations Diversity factor data between alternative symptom relevant information, comprising: the posterior probability values maximum and the default symptom taken second place are related One group of target association degree that information corresponds to each default symptom information is not less than zero difference between.
In one embodiment of the invention, the medicine belongs to traditional Chinese medical science field, the relevant information library include: comprising one or The card type information bank of multiple card type information, and/or prescription information library comprising one or more prescription informations.
It include main prescription information in the prescription information in one embodiment of the invention；The main prescription information is used for Substitute the prescription information.
In one embodiment of the invention, the processing unit is also used to by distributed weight from the association symptom sequence The maximum data of quantity magnitude of corresponding respective weight are chosen in column and one or more described project identification symptom sequences respectively, And the corresponding default symptom information of those data is gathered to form comprehensive identification symptom sequence.
In one embodiment of the invention, the processing unit is also used to adjust the weight according to interrogation phase difference.
In order to achieve the above objects and other related objects, the present invention provides a kind of medical data recommended method, is applied to one It is stored with the processing unit of structured patient record data set；Wherein, the structured patient record data set includes at least one structuring Project information under medical record data, each structured patient record data and its various classifications of the items for being included is to a medicine Case history concentrate a case history and its include each project content mathematical character；The classification of the items includes: comprising one Or Symptomatic classification and the relative symptom relevant classification comprising one or more symptom relevant informations of multiple symptom informations； The processing unit is also stored with: gathering what various unduplicated default symptom informations in the structured patient record data set were formed It is various in default symptom information library and the medicine structure data set that the various classifications of the items are gathered respectively not repeat Default symptom relevant information formed each relevant information library；The processing unit is also stored with: the default symptom information Between each default symptom information in library, and/or in each default symptom information and relevant information library between each default symptom relevant information Degree of association data；Institute's method includes: to obtain the symptom information sequence to be diagnosed comprising one group of symptom information to be diagnosed；In symptom In degree of association data in information bank between each default symptom information, matches and calculate one group of symptom information to be diagnosed point Degree of association data not between each default symptom information, set form association symptom sequence, for choosing the wherein degree of association Default symptom information corresponding to one or more maximum data of the magnitude of data is as recommendation symptom information；And/or with The structured patient record data set is inputted as data, calculates the relevant information library occurred based on each symptom information to be diagnosed Each of the posterior probability that occurs of default symptom relevant information；Posterior probability values are selected in the respectively default symptom relevant information Maximum one as recommendation symptom relevant information；Alternatively, the maximum multiple default symptom correlations of posterior probability values are believed Alternately symptom relevant information is ceased, it is default to judge whether the magnitude of the proximity data between each alternative symptom relevant information is lower than Value；If it is not, then selecting the maximum alternative symptom relevant information of posterior probability values as recommendation symptom relevant information；If so, Degree of association data in the symptom information library in each default symptom information and relevant information library between each default symptom relevant information In, match each alternative symptom relevant information target association degree evidence between each default symptom information respectively, and according to being matched to Target association degree according to calculating: be relevant to the alternative of default symptom informations all each of different from each symptom information to be diagnosed Diversity factor data between symptom relevant information, set form project and distinguish symptom sequence, for choosing wherein diversity factor data One or more maximum data of magnitude corresponding to default symptom information as recommend symptom information as recommending symptom information.
In order to achieve the above objects and other related objects, the present invention provides a kind of processing unit, including processor and storage Device；The memory, for store the first computer program, the processor, for run first computer program with Realize the medical history data processing method；Alternatively, the memory, for storing by the medical history data The medical structure medical record data collection that reason method generates；Alternatively, the memory, for storing second computer program；It is described Processor realizes the medical data recommender system for running the second computer program.
In order to achieve the above objects and other related objects, the present invention provides a kind of computer storage medium, the first meter of storage Calculation machine program, first computer program are used to run by processor to realize the medical history data processing method； Alternatively, the medical structure medical record data collection that storage is generated by the medical history data processing method；Alternatively, storage second Computer program, the second computer program are used to run by processor to realize the medical data recommender system.
As described above, medical history data processing method of the invention, medical data recommender system, equipment and medium, institute State medical history data processing method include: by a medical history concentrate each case history and its include each project content It is expressed as the project information under a structured patient record data and its various classifications of the items for being included by mathematical character mode, To form the medical structure medical record data collection for including one or more structured patient record data；In turn, medical data is recommended System can be carried out based on the medical structure medical record data collection to the data analysis about information such as symptoms, to predict and recommend Meet the interrogation foundation in efficient interrogation direction, auxiliary doctor of traditional Chinese medicine realizes efficient interrogation process, the realistic meaning with height.
Detailed description of the invention
Fig. 1 is shown as the flow diagram of traditional Chinese medicine medical record data processing method of the embodiment of the present invention.
Fig. 2 is shown as extracting the step of the degree of association data in symptom information library between various symptoms information in the embodiment of the present invention Rapid flow chart.
Fig. 3 is shown as extracting in symptom information library symptom in various symptoms information and relevant information library in the embodiment of the present invention The step flow chart of degree of association data between relevant information
Fig. 4 is shown as the module diagram of traditional Chinese medicine data recommendation system of the embodiment of the present invention.
Fig. 5 is shown as obtaining association symptom sequence in the embodiment of the present invention and recommends the step flow chart of symptom information.
Fig. 6 A is shown as obtaining the recommendation symptom correlation letter of corresponding symptom information sequence to be diagnosed in one embodiment of the invention The step flow diagram of breath.
The recommendation symptom that Fig. 6 B is shown as obtaining corresponding symptom information sequence to be diagnosed in further embodiment of this invention is related The step flow diagram of information.
Specific embodiment
Illustrate embodiments of the present invention below by way of specific specific example, those skilled in the art can be by this specification Other advantages and efficacy of the present invention can be easily understood for disclosed content.The present invention can also pass through in addition different specific realities The mode of applying is embodied or practiced, the various details in this specification can also based on different viewpoints and application, without departing from Various modifications or alterations are carried out under spirit of the invention.It should be noted that in the absence of conflict, following embodiment and implementation Feature in example can be combined with each other.
It should be noted that illustrating the basic structure that only the invention is illustrated in a schematic way provided in following embodiment Think, only shown in schema then with related component in the present invention rather than component count, shape and size when according to actual implementation Draw, when actual implementation kenel, quantity and the ratio of each component can arbitrarily change for one kind, and its assembly layout kenel It is likely more complexity.
Firstly, there is a problem of that the present invention adds it in a jumble without science based on medical history data in the prior art To improve.Specifically, the statement of above-mentioned medicine, which can be, belongs to traditional Chinese medical science field, mainly closes in subsequent embodiment In the citing of the technical solution of traditional Chinese medical science field, but this is not to limit the present invention to can be only applied to traditional Chinese medical science field；According to subsequent interior Hold it is found that the present invention can also apply in doctor trained in Western medicine field.
As shown in Figure 1, showing the flow diagram of traditional Chinese medicine of embodiment of the present invention medical record data processing method.
The described method includes:
Step S101: by each case history that a medical history is concentrated and its including that the content of each project passes through mathematical table Sign mode is expressed as the project information under a structured patient record data and its various classifications of the items for being included.
Specifically, Chinese medicine medical record data generally includes the following contents: patient basis, sign, symptom, previously disease The projects such as history, card type and prescription.
For example, the essential information includes such as gender, age, height, weight；The sign includes body temperature, arteries and veins It fights, breathe, blood pressure etc., symptom is also known as subjective symptoms, refers to the master of the patient in morbid state lower body physiological function exception Perception is by such as shortness of breath, pectoralgia, vomiting etc.；Certainly, some symptoms are also sign, for example are wheezed, and patient oneself is described as a moment A burst of wheezes, this is a symptom；Doctor can also hear whoop in physical examination, therefore be also sign, so partial symptoms Be with sign it is identical, the part sign and symptom can be unified.
The previous sufferer record of the medical history, that is, patient.
The card type is a kind of title specific to Chinese medicine.Card, both syndrome, referred to the disease of certain phase in lysis Position, the cause of disease, characteristic of disease, patient's condition and body resistance against diseases the essential organic connections such as power reactiveness, showing as clinic can quilt The symptom etc. observed.
For example, summer red swelling of the skin malignant boil furuncle occurs because heated, exactly the variation (blood-head) of blood is caused to cause by the cause of disease (heat) The morbid state (malignant boil furuncle, blood-heat syndrome) of human body；For another example by flu for, have chill and wind-heat and summer-wet type etc., wherein chill and Wind-heat is exactly the summary to the attribute of this disease, that is, " card type ".
That is, card type and symptom are closely bound up.
Prescription is that the Traditional Chinese medicinal prescription suited the remedy to the case outputed after card type is determined according to symptom in Chinese medicine, is also and symptom Closely bound up.
Since traditional Chinese medicine case history is usually the unstructured case history write with natural language, it is unfavorable for carrying out data system Meter analysis；Thus, the improvement of the application first is that handling unstructured case history for structured patient record.
For example, a method of handling unstructured case history for structured patient record are as follows: extract in original case history Information element carries out mathematical way (such as coding staff in the respective item classification in structured patient record data with independent field Formula) it is characterized as project information, it is stored later, that is, forms structured patient record data.
The classification of the items include: the Symptomatic classification comprising one or more symptom informations and it is relative include one or The symptom relevant classification of multiple symptom relevant informations, for example, syndrome-classification, prescription classification etc.；It is every in structuring case data A classification of the items exists with the independent data part marked off, wherein storage is attributed to the one or more of project information of this classification.
For example, there is headache information in the patient symptom with natural language description in original case history, then by headache information It extracts and is stored in structured patient record data under Symptomatic classification as a symptom information.
The structure of bright structuring case data in further example, in original case history after above structureization processing, Structured patient record data with the following characteristics can be formed, structure includes: the base of one group of patient under essential information classification This information G={ g1, g2 ..., gn }, the symptom information sequence PS={ ps1, ps2 ..., psn } to be diagnosed under Symptomatic classification, Card type information D under syndrome-classification, the prescription information P under prescription classification.
It is noted that above-mentioned classification of the items is not limited to traditional literal meaning, it might even be possible to by multiple items Comprehensive mesh classification is one, or is subject to cutting, and or is substituted etc. with the sub-project classification in classification of the items to it.
For example, in one or more embodiments, " symptom " classification also can have broader meaning, can also To include the sign information of patient, that is to say, that can unify Symptomatic classification and sign classification for Symptomatic classification；On in addition, One group of main prescription information F and one group of plus-minus traditional Chinese medicine ingredients information M can be further represented as by stating prescription again, carry out actual statistics When analysis, it can be calculated by main prescription information F to substitute prescription information.
According to the example above, the structure of every structured patient record data is represented by (G, PS, D, P (F, M)).
Step S102: the medical structure medical record data collection comprising one or more structured patient record data is formed.
It is formed since every structured patient record data can be a corresponding case history, then card type information and prescription information can With only one；Corresponding structuring can be converted to by the medical history collection that one or more unstructured medical record datas form Medical record data collection.
For example, structured patient record data set R={ R1,R2,…,RnIt altogether include n parts of structured patient record data, wherein I-th part of case history RiIt is represented by (Gi,PSi,Di,Pi(Fi,Mi))。
It, can be between the medicine correlation analysis various project information, with benefit after getting structured patient record data set The medical data needed for therefrom effective acquisition, especially with respect between symptom information, between symptom information and prescription information and symptom believe The analysis of degree of association data between breath and card type information.
For this purpose, various unduplicated symptom informations can be extracted first from the medicine structure data set, storage forms disease Shape information bank, for example, assuming that occurring m mutually different symptom { s in this group of medical record data in total1,s2,…,sm, The symptom library of referred to as this group medical record data；
Alternatively, it is also possible to carry out building library for the symptom relevant classification more paid close attention to；It specifically, can be from the structure Changing medical record data concentrates corresponding each symptom relevant classification to extract various unduplicated symptom relevant informations respectively, storage formed with The corresponding relevant information library of each symptom relevant classification.
For example, the symptom relevant classification such as syndrome-classification and prescription classification etc., can be from structured patient record number Unduplicated each card type information correspondence establishment card type information bank is extracted according to concentrating, it is assumed for example that structured patient record data set R= {R1,R2,…,RnIn occur m mutually different symptom information { s in total1,s2,…,sm, then { s1,s2,…,smBe referred to as For the symptom information library of this group of medical record data；Unduplicated each prescription information correspondence establishment is extracted from structured patient record data set Prescription information library, it is assumed for example that this group of structured patient record data R={ R1,R2,…,RnIn to occur m in total mutually different Prescription information { p1,p2,…,pm, then { p1,p2,…,pmIt is referred to as the prescription information library of this group of medical record data.
Further alternative, the medical data processing method, which may also include that, carries out data according to structure medical record data collection Analysis is to extract the degree of association data in above-mentioned various project information libraries between project information.
For example, it can extract between each symptom information in symptom information library, and/or each symptom information and relevant information Degree of association data in library between each symptom relevant information, and stored.
As shown in Fig. 2, showing the step flow chart for extracting the degree of association data in symptom information library between various symptoms information.
The process includes:
Step S201: every kind of symptom information in corresponding symptom information library establishes symptom and vector occurs；Wherein, institute It states symptom and the set that vector is expressed as the characterization value whether every kind of symptom information appears in each structured patient record data occurs.
For example, it is based on this group of medical record data, the symptom that m dimension can be established for each symptom sj in symptom library goes out Existing vector (r1, r2 ..., rn), wherein ri represents whether symptom sj appears in the patient symptom information PSi of case history Ri.If Symptom sj is appeared in the patient symptom information PSi of case history Ri.Then ri=1, otherwise ri=0.
Step S202: it calculates symptom corresponding to various symptoms information and the similarity information between vector occurs, as the pass Join degree evidence.
In one embodiment of the invention, the calculation method of the similarity information include: COS distance, Euclidean distance, Standardize any one in Euclidean distance, mahalanobis distance, Hamming distance and manhatton distance.
It is illustrated by taking COS distance calculation as an example, calculates in the symptom library of this group of medical record data all symptoms two-by-two Between symptom there is the COS distance of vector.With Y (si,sj) represent symptom siWith symptom sjSymptom there is the cosine of vector Distance, and the similarity between symptom, the referred to as similarity information of symptom are represented with the COS distance value.Two symptoms it is similar Degree information height can represent that the probability that two symptoms appear in the same case history is higher, and also representing two symptoms may have to a certain degree There is certain medicine relevance.
The set of the cosine similarity of all symptoms between any two, which then constitutes, in the symptom library of this group of medical record data is based on The symptom association table of this group of medical record data is as shown in table 1 below, and is stored.
Table 1
Symptom s1 Symptom s2 。。。 Symptom si 。。。 Symptom sm
Symptom s1 Null Y(s2,s1) 。。。 Y(si,s1) 。。。 Y(sm,s1)
Symptom s2 Y(s1,s2) Null 。。。 Y(si,s2) 。。。 Y(sm,s2)
。。。 。。。 。。。 。。。 。。。 。。。 。。。
Symptom sj Y(s1,sj) Y(s2,sj) 。。。 Y(si,sj) 。。。 Y(sm,sj)
。。 。。。 。。。 。。。 。。。 。。。 。。。
Symptom sm Y(s1,sm) Y(s2,sm) 。。。 Y(si,sm) 。。。 Null
In one or more embodiments of the invention, further can also in symptom information bank symptom information to it is related The degree of association data between symptom relevant information in information bank are extracted, such as extract the association of symptom information and card type information Degree evidence, degree of association data between symptom information and prescription information etc. can be presented by way of table.
Various symptoms information and relevant information library in symptom information library are extracted in the embodiment of the present invention as shown in figure 3, showing The step flow chart of degree of association data between middle symptom relevant information.
Step S301: counting respectively in the case of each symptom relevant information occurs in relevant information library and symptom information The quantity for the structured patient record data that various symptoms information occurs in library accounts for the ratio of structured patient record data set respectively；
Step S302: being to occur every kind of symptom letter in the case of every kind of symptom relevant information occurs by the ratio approximate representation The conditional probability of breath, as the degree of association data.
That is, the degree of association data, which can be, is formed by conditional probability in such a way that conditional probability is analyzed Value.
It is illustrated with the degree of association data between card type information and symptom information, such as structured patient record data set R= { R1, R2 ..., Rn } altogether include n parts of structuring case data, wherein i-th part of structuring case data Ri be represented by (Gi, PSi,Di,Pi(Fi,Mi)).Wherein PSi is the Symptomatic classification that the structuring case data include, and it includes the collection of symptom information It closes, Di is the syndrome-classification that the Symptomatic classification that the structuring case data include includes, and it includes card type information.Assuming that total in R Occur k mutually different card type information 1, d2 ..., dk altogether }, the card type information bank of referred to as R；And m is a mutually different Symptom information { s1, s2 ..., sm }, the symptom library of referred to as R.Based on structured patient record data set R, can count in certain card type Information dj occur case in certain symptom information si simultaneously occur structuring case data quantity and account for structuring case in R The ratio of data bulk total quantity (n).When the data volume of the R is sufficiently large (such as n is sufficiently large), in the case history that certain card type dj occurs In certain symptom si simultaneously occur ratio, can approximation representative certificate type information dj occur under symptom si information appearance condition it is general Rate: P (si | dj).When the conditional probability value is higher, according to the degree of association of Chinese medicine common sense symptom information si and card type information dj Also higher.
Accordingly, all symptom informations in the symptom information library of structured patient record data set R can be calculated and be based on card The set of each conditional probability P (si | dj) of all card type information, forms the disease based on card type information of R in type information bank The conditional probability table of shape information, as shown in table 2 below.
Table 2
Symptom s1 Symptom s2 。。。 Symptom sj 。。。 Symptom sm
Card type d1 P(s1|d1) P(s2|d1) 。。。 P(sj|d1) 。。。 P(sm|d1)
Card type d2 P(s1|d2) P(s2|d2) 。。。 P(sj|d2) 。。。 P(sm|d2)
。。。 。。。 。。。 。。。 。。。 。。。 。。。
Card type dj P(s1|dj) P(s2|dj) 。。。 P(sj|dj) 。。。 P(sm|dj)
。。 。。。 。。。 。。。 。。。 。。。 。。。
Card type dk P(s1|dk) P(s2|dk) 。。。 P(sj|dk) 。。。 P(sm|dk)
It is illustrated again with the degree of association data between prescription information and symptom information, such as structured patient record data set R= { R1, R2 ..., Rn } altogether include n parts of structuring case data, wherein i-th part of structuring case data Ri be represented by (Gi, PSi,Di,Pi(Fi,Mi)).Wherein PSi is the Symptomatic classification that the structuring case data include, and it includes the collection of symptom information It closes,
It should be noted that in the present embodiment, using preferred mode, by main prescription information F substitute prescription information P come It is associated the calculating of degree evidence.Certainly, it and non-limiting cannot directly be calculated with prescription information P.
Fi is the main prescription information that the case history includes.Assuming that occur in total in R k mutually different main prescription f1, F2 ..., fk }, the main prescription information bank of referred to as R；And m mutually different symptoms { s1, s2 ..., sm }, the referred to as symptom of R Library.Based on R, the structuring case load certain symptom letter si in the case history that certain main prescription information fj occurs while occurred can be counted According to quantity and account for the ratio of structuring case data bulk total quantity (n) in R.(such as n foot when the data volume of the R is sufficiently large It is enough big), the ratio that certain symptom information si occurs simultaneously in the case history that certain main prescription information fj occurs, can approximation represent main prescription The conditional probability that symptom information si under information fj appearance occurs: P (si | fj).It is normal according to Chinese medicine when the conditional probability value is higher Symptom information si and the degree of association of main prescription information fj are also higher known to knowing.
Conditional probability of all symptom informations based on main prescription information all in main prescription information bank in the symptom information library of R The set of P (si | fj), forms the conditional probability table of the symptom information based on main prescription information of R, as shown in table 3 below.
Table 3
Symptom s1 Symptom s2 。。。 Symptom sj 。。。 Symptom sm
Main prescription f1 P(s1|f1) P(s2|f1) 。。。 P(sj|f1) 。。。 P(sm|f1)
Main prescription f2 P(s1f2) P(s2|f2) 。。。 P(sj|f2) 。。。 P(sm|f2)
。。。 。。。 。。。 。。。 。。。 。。。 。。。
Main prescription fj P(s1|fj) P(s2|fj) 。。。 P(sj|fj) 。。。 P(sm|fj)
。。 。。。 。。。 。。。 。。。 。。。 。。。
Main prescription fk P(s1|fk) P(s2|fk) 。。。 P(sj|fk) 。。。 P(sm|fk)
After the contingency table for obtaining above-mentioned expression degree of association data, it can carry out further statisticalling analyze to predict energy Next interrogation target of high efficiency acquisition diagnostic result.
For example, during interrogation, after determining a symptom information, it is next how doctor chooses for doctor A symptom information progress inquiry, can be by above structure case history efficiently can quickly determine the information such as card type, prescription Data set and table 1, table 2, table 3 etc. are for statistical analysis to be obtained.
For this purpose, as shown in figure 4, showing the module signal of medical data recommender system provided by the invention Figure.The system is applied to a processing unit.
The processing unit can be the electronic equipment with data-handling capacity, such as computer, smart phone, plate Computer etc.；The processing unit is also possible to multiple electronic equipments with data-handling capacity and is connected to the network and is cooperated It is formed by network processing device, the network includes Internet of Things, internet, Intranet, wide area network (WAN), local area network (LAN), wireless network, Digital Subscriber Line (DSL) network, frame-relay network, asynchronous transfer mode (ATM) network, virtual private Any one or more of network (VPN) and/or any other suitable communication network；The processing unit is also possible to one A electronic component with data-handling capacity and it is integrated in the electronic device, the electronic component includes processor and storage Device, computer program in processor run memory and realize function.
Degree of association data, project information library between the structured patient record data set obtained in advance, project information etc. are equal It can store in the memory of the processing unit for using, in the examples below, by pre-stored data, letter Increase the statement, such as " default symptom information " etc. of " default " before breath.
Institute's system includes: receiving unit 401 and processing unit 402.
In one embodiment of this invention, which can statistically analyze must recommend according to degree of association data between symptom information Symptom information
The receiving unit 401, for obtaining the symptom information sequence to be diagnosed comprising one group of symptom information to be diagnosed.
For example, during clinical interrogation, when user has inputted one group of symptom information to be diagnosed, including k Mutually different symptom information to be diagnosed forms symptom information sequence PS={ ps to be diagnosed1, ps2..., psk}。
The processing unit 402, in the degree of association data between the default symptom information each in symptom information library, Degree of association data of the one group of symptom information to be diagnosed respectively between each default symptom information are matched and calculated, shape is gathered At association symptom sequence, for choosing default symptom corresponding to wherein one or more maximum data of magnitude of degree of association data Information is as recommendation symptom information.
Specifically, in one embodiment, as shown in figure 5, showing acquisition association symptom sequence and recommending symptom information Flow chart of steps.
As shown, the process specifically includes:
Step S501: it in the degree of association data between default symptom information, matches each described to diagnostic symptom and each Current degree of association data between the default symptom information.
For example, it to each of PS={ ps1, ps2 ..., psk } symptom psj, is formed by inquiring above procedure Symptom association table (i.e. table 1), match the similarity in symptom information psj and symptom information library between other symptoms information and believe Breath, and will be similar to symptom information default in symptom information library by all symptom informations in the PS={ ps1, ps2 ..., psk } It spends information and inserts current symptomatic association table below, as shown in table 4 below.
In one embodiment of the invention, the calculation method of the similarity information include: COS distance, Euclidean distance, Standardize any one in Euclidean distance, mahalanobis distance, Hamming distance and manhatton distance.
Step S502: respectively obtaining summed result to each current degree of association data relevant to each default symptom information, The association symptom sequence is constructed with the set of each summed result；Wherein, in the association symptom sequence, the summation knot The magnitude of fruit is bigger, corresponding to default symptom information as recommend symptom information priority it is higher.
For example, in table 4, each default symptom information can be corresponded in turn by itself and current symptomatic association table Each column wait diagnose the summation of the degree of association data between symptom information, group symptom information to be diagnosed and symptom are represented with this and believed Cease the degree of association between the default symptom of other in library.
Table 4
The summed result that collection closes last line in table forms association symptom sequence.Since the magnitude of degree of association data is got over Greatly, indicate that the degree of association is higher, it therefore, preferably can be by the magnitude of each summed result from big in the degree of association symptom sequence It is ranked up to small, in order of sequence the big preferential output of degree of being associated.
In practice, a certain number of symptom informations, example can be chosen in association symptom sequence by certain way Such as, symptom information corresponding to maximum top n symptom or the degree of association data of the big Mr. Yu's particular value of magnitude etc. is chosen, as The recommendation symptom information of next step interrogation foundation during Chinese medicinal interrogation.
In one embodiment of this invention, the processing unit 402 can be used for selection recommendation symptom relevant information work For next step interrogation foundation.
As shown in Fig. 6 A and 6B, the recommendation symptom phase that corresponding symptom information sequence to be diagnosed is obtained in two embodiments is shown Close the step flow diagram of information.Symptom relevant information as previously described includes such as card type information, prescription information.
The process specifically includes:
Step S601: being inputted using the structured patient record data set as data, is calculated based on each described to diagnostic symptom The posterior probability that the default symptom relevant information in each of relevant information library that information occurs occurs.
In one embodiment of the invention, the calculation of the posterior probability can be the calculation based on Bayesian formula Method, such as NB Algorithm and its mutation algorithm.
Step S602: select in the respectively default symptom relevant information posterior probability values maximum one as recommendation symptom Relevant information.
In one embodiment of the invention, fairly simple mode can be directly with the posterior probability values maximum one As symptom relevant information is recommended, for example, choosing the maximum card type information of posterior probability values as target card type information, after selection The maximum prescription information of probability value is tested as target prescription information etc..
Diversity factor but in a kind of symptom relevant information, between the maximum multiple symptom relevant informations of posterior probability values (can be indicated by the absolute value of the difference of corresponding posterior probability values) is smaller, then it is maximum to choose posterior probability values Symptom relevant information is more unreliable as the way of recommendation symptom relevant information.
Therefore, it can improve this, as shown in Figure 6B, by following steps come alternative steps S602.
Step S603: by the maximum multiple default symptom relevant informations of posterior probability values, alternately symptom correlation is believed Breath, judges whether the magnitude of the proximity data between each alternative symptom relevant information is lower than preset value.
Step S604: if it is not, then selecting the maximum alternative symptom relevant information of posterior probability values related as recommendation symptom Information.
Specifically, in the maximum multiple alternative symptom relevant informations of posterior probability values, mutual posterior probability values Difference it is bigger, then diversity factor is bigger, correspondingly, maximum alternative symptom relevant information as recommend symptom relevant information can It is higher by spending.
Step S605: if so, each default symptom information is preset with each in relevant information library in the symptom information library In degree of association data between symptom relevant information, each alternative symptom relevant information mesh between each default symptom information respectively is matched Mark degree of association data；
Step S606: according to the target association degree being matched to according to calculating: being relevant to each symptom information to be diagnosed all Diversity factor data between the alternative symptom relevant information of each of different default symptom informations, set form project and distinguish symptom Sequence pushes away so that default symptom information corresponding to selection wherein one or more maximum data of magnitude of diversity factor data is used as Symptom information is recommended as recommendation symptom information.
In one embodiment of the invention, the multiple alternative symptom relevant information includes: that posterior probability values are maximum and secondary Default symptom relevant information, that is, choose maximum two default symptom relevant informations；It is described be relevant to it is each to Diversity factor data between the alternative symptom relevant information of all each of different default symptom informations of diagnostic symptom information, comprising: One group of target that the posterior probability values maximum and the default symptom relevant information taken second place correspond to each default symptom information is closed Join the difference (i.e. absolute value) that degree is not less than zero between.
It is illustrated with card type information, some in verification type information bank { d1, d2 ..., dk } presets card type information dj, the card Type information be based on should symptom information sequence PS=be diagnosed { ps1, ps2 ..., psk } posterior probability be represented by P (dj | ps1, ps2,…,psk).Using model-naive Bayesian:
P(dj|ps1,ps2,…,psk)
∝P(ps1,ps2,…,psk|dj)*P(dj)
∝P(ps1|dj)*P(ps2|dj)*…*P(psk|dj)*P(dj)；
All default card type information in card type information bank { d1, d2 ..., dk } are calculated based on group symptom information to be diagnosed Posterior probability, and all card type information are ranked up from big to small according to posterior probability values；If posterior probability values highest two Probability value difference between a alternative card type information is greater than certain predetermined value, then can be by posterior probability values highest one alternative card Type information is as the target card type information based on symptom information sequence to be diagnosed, to be provided as next step interrogation foundation.
And if the probability value gap between the highest two alternative card type information of sorting is less than certain predetermined value, needs pair This two alternative card type information are further recognized.
Assuming that two alternative card type information for needing further to recognize are d1 and d2, then symptom information library is obtained from table 2 In all default symptom informations be based respectively on the conditional probability value of alternative card type information d1, d2, and calculate its difference (such as difference be It is negative then take absolute value), following card type identification symptom table is inserted, i.e., as shown in table 5.
It will be excluded in symptom information library each described wait diagnose all default symptom informations other than symptom information according to above The size of the conditional probability difference of calculating is ranked up, and is formed card type and is recognized symptom sequence.Conditional probability difference is bigger, then represents The symptom information and the degree of association of one of card type information are bigger, and smaller with the degree of association of another card type information；? The information that patient's symptom is obtained during subsequent interrogation then facilitates to differentiate card type letter in two alternative card type information Breath.
Table 5
Symptom s1 。。。 Symptom si 。。。 Symptom sm
Card type d1 P(s1|d1) 。。。 P(si|d1) 。。。 P(sm|d1)
Card type d2 P(s1|d2) 。。。 P(si|d2) 。。。 P(sm|d2)
Probability difference |P(s1|d1)-P(s1|d2)| 。。。 |P(si|d1)-P(si|d2)| 。。。 |P(sm|d1)-P(sm|d2)|
During Chinese medicinal interrogation, after doctor obtains certain amount symptom information, by analysis to existing information and Thinking, doctor would generally make preliminary diagnosis, form alternative diagnostic result, need to obtain at this time additional symptom information with It is further to diagnosis to be confirmed.The generation type of this system card type identification symptom sequence meets the right mind mistake of doctor of traditional Chinese medicine Journey, i.e., by analyzing one group of symptom information to be diagnosed, the foundation that being formed helps further to recognize card type recommends symptom The interrogation indication of information progress interrogation.
Similarly, then by taking prescription information as an example the explanation of target prescription information formation is carried out, uses main prescription in the present embodiment Information substitutes prescription information.
To some main prescription information fj in main prescription information bank { f1, f2 ..., fk }, which is based on the follow-up The posterior probability of disconnected symptom information sequence PS={ ps1, ps2 ..., psk } be represented by P (fj | ps1, ps2 ..., psk).Using Model-naive Bayesian:
P(fj|ps1,ps2,…,psk)
∝P(ps1,ps2,…,psk|fj)*P(fj)
∝P(ps1|fj)*P(ps2|fj)*…*P(psk|fj)*P(fj)
All main prescription information in main prescription information bank { f1, f2 ..., fk } are calculated based on should symptom information sequence be diagnosed Posterior probability, and all main prescription information are ranked up according to probability value；If sorting highest two alternative main prescription letters Probability value difference between breath is greater than certain predetermined value, then can using posterior probability values highest one alternative main prescription information as The main prescription information of target based on the symptom information sequence to be diagnosed.
And if the probability value gap between the highest two alternative main prescription information that sorts is less than certain predetermined value, needs Two alternative main prescriptions are further recognized.
Assuming that two alternative main prescription information for needing further to recognize are f1 and f2, then symptom information is obtained from table 3 All default symptom informations are based on the conditional probability value of alternative main prescription information f1, f2 in library, and calculate its difference (such as difference is It is negative then take absolute value), the following main prescription of filling recognizes symptom table, i.e., as shown in table 6.
It will be excluded in symptom information library each described wait diagnose all default symptom informations other than symptom information according to above The size of the conditional probability difference of calculating is ranked up, and forms main prescription identification symptom sequence.Conditional probability difference is bigger, then generation The table symptom and the degree of association of one of them main prescription are bigger, and smaller with the degree of association of another main prescription；Subsequent The information that patient's symptom is obtained during interrogation then facilitates the main prescription letter of the resolution target in two alternative main prescription information Breath.
Table 6
Symptom s1 。。。 Symptom si 。。。 Symptom sm
Main prescription f1 P(s1|f1) 。。。 P(si|f1) 。。。 P(sm|f1)
Main prescription f2 P(s1|f2) 。。。 P(si|f2) 。。。 P(sm|f2)
Probability difference |P(s1|f1)-P(s1|f2)| 。。。 |P(si|f1)-P(si|f2)| 。。。 |P(sm|f1)-P(sm|f2)|
During Chinese medicinal interrogation, after doctor obtains certain amount symptom information, by analysis to existing information and Thinking, doctor would generally make preliminary diagnosis, form alternative diagnostic result, need to obtain at this time additional symptom information with It is further to diagnosis to be confirmed.The generation type of the main prescription identification symptom sequence of this system meets the right mind of doctor of traditional Chinese medicine Process forms the foundation recommendation for helping further to recognize main prescription that is, by analyzing one group of symptom information to be diagnosed The interrogation indication of symptom information progress interrogation.
In one embodiment of the invention, symptom sequence is recognized from the association symptom sequence and one or more described projects It can be used alone, to provide next step interrogation foundation.It for example, can be in association symptom sequence, card type identification sequence or master Recommendation symptom information corresponding to one or more the maximum data of magnitude selected in prescription identification symptom sequence is directly output to Doctor, using the foundation as progress next step interrogation.
However, it is preferred that being exported again after those sequences being integrated, using the foundation as progress next step interrogation.
In one embodiment of this invention, the processing unit 402, it may also be used for press distributed weight from the associated diseases The quantity magnitude for choosing corresponding respective weight in shape sequence and one or more described project identification symptom sequences respectively is maximum Data, and the corresponding default symptom information of those data is gathered to form comprehensive identification symptom sequence.
For example, when obtained by above method the association symptom sequence based on the symptom information sequence to be diagnosed, Card type recognize symptom sequence and main prescription (or prescription) identification symptom sequence after, can using certain method to three sequences into Row integration, forms final interrogation indication symptom sequence.A kind of method of feasible integration is that give each sequence certain Weight, and respectively take a certain number of symptom informations from each sequence according to the ratio of weight.
For example, have 10 values in association symptom sequence, corresponding 10 default symptom informations, and its weight is 0.2；If follow-up Disconnected symptom information has 4 in 10 default symptom informations, then in card type identification symptom sequence and main prescription identification symptom sequence There are 6 values, corresponding remaining 6 default symptom informations, and the weight for setting card type identification sequence is 0.3, main prescription identification disease The weight of shape sequence is 0.4.
2 corresponding default symptom informations of the maximum 10*0.2=2 value of magnitude are chosen from association symptom sequence, from Card type, which recognizes, chooses 2 corresponding default symptom informations of maximum 6*0.3=1.8 ≈ 2 values of magnitude in symptom sequence, from master Prescription, which recognizes, chooses 1 corresponding default symptom information of maximum 6*0.2=1.2 ≈ 1 value of magnitude in symptom sequence, then should 5 select the default symptom information set come as comprehensive identification symptom sequence, and then export, and can be supplied to doctor's behaviours The foundation of further interrogation.
However, a kind of mode more optimized is then that can be dynamically adapted according to the different phase of interrogation process each The weight of sequence；For example, doctor mainly needs to obtain patient's symptom information complete as far as possible, this rank at the initial stage of interrogation process Section according to the degree of association between symptom information guide prompt it is even more important, then higher weight can be distributed by being associated with symptom sequence； In the later period of interrogation process, then even more important for the discrimination of alternatively card type information and main prescription information, card type recognizes disease at this time Shape sequence and main prescription identification symptom sequence should distribute higher weight.
Output as the system is supplied to doctor by the interrogation indication symptom sequence after integration, and doctor refers to the letter Breath carries out the interrogation of next step, it will help doctor realizes efficient interrogation process.
It should be noted that although refer to each sequence carrying out sequence from big to small in above-described embodiment, this be in order to Convenient for the direct output carried out in order of sequence from big to small in computer equipment；Certainly, this is a kind of preferred embodiment, real May not necessarily also sort on border, every time carry out maximum value calculation also can, be not limited thereto.
It should be noted that it should be understood that the division of each unit in system above is only a kind of drawing for logic function Point, it can completely or partially be integrated on a physical entity in actual implementation, it can also be physically separate.And these units can All to be realized by way of processing element calls with software；The shape of software can be called by processing element with unit Formula realizes that unit passes through formal implementation of hardware.For example, processing unit 402 can be the processing element individually set up, It can integrate and realized in some chip of above-mentioned apparatus, in addition it is also possible to be stored in above-mentioned dress in the form of program code In the memory set, the function of the above processing unit 402 is called by some processing element of above-mentioned apparatus and executed.Other lists The realization of member is similar therewith.Furthermore these units completely or partially can integrate together, can also independently realize.It is described here Processing element can be a kind of integrated circuit, the processing capacity comprising signal.During realization, above steps or more Each unit can be completed by the integrated logic circuit of the hardware in processor elements or the instruction of software form.
For example, the above unit can be arranged to implement one or more integrated circuits of above method, such as: One or more specific integrated circuits (ApplicationSpecificIntegratedCircuit, abbreviation ASIC), or, one Or multi-microprocessor (digitalsingnalprocessor, abbreviation DSP), or, one or more field-programmable gate array Arrange (FieldProgrammableGateArray, abbreviation FPGA) etc..For another example, when some above unit is dispatched by processing element When the form of program code is realized, which can be general processor, such as central processing unit (CentralProcessingUnit, abbreviation CPU) or it is other can be with the processor of caller code.For another example, these units can To integrate, realized in the form of system on chip (system-on-a-chip, abbreviation SOC).
In one embodiment of the present of invention, the realization of corresponding above system, the present invention can also provide medical data recommendation Method, institute's method include: to obtain the symptom information sequence to be diagnosed comprising one group of symptom information to be diagnosed；In symptom information library In degree of association data between each default symptom information, match and calculate one group of symptom information to be diagnosed respectively with it is described Degree of association data between each default symptom information, set form association symptom sequence, for choosing the amount of wherein degree of association data It is worth default symptom information corresponding to one or more maximum data as recommendation symptom information；And/or with the structure Change medical record data collection to input as data, calculates each of the relevant information library occurred based on each symptom information to be diagnosed The posterior probability that default symptom relevant information occurs；Posterior probability values maximum one are selected in the respectively default symptom relevant information A be used as recommends symptom relevant information；Alternatively, using the maximum multiple default symptom relevant informations of posterior probability values as standby Symptom relevant information is selected, judges whether the magnitude of the proximity data between each alternative symptom relevant information is lower than preset value；If it is not, Then select the maximum alternative symptom relevant information of posterior probability values as recommendation symptom relevant information；If so, in the symptom In degree of association data in information bank in each default symptom information and relevant information library between each default symptom relevant information, matching is each The alternative symptom relevant information target association degree evidence between each default symptom information respectively, and according to the target association being matched to Degree is according to calculating: being relevant to the related letter of all each of different to each symptom information to be diagnosed default alternative symptoms of symptom information Diversity factor data between breath, set form project and distinguish symptom sequence, so that the magnitude for choosing wherein diversity factor data is maximum One or more data corresponding to default symptom information as recommend symptom information as recommending symptom information.
Since the principle of the method is roughly the same with above system, the technical characteristic in an embodiment all be can be universally used in separately One embodiment, thus for it has been described that technical detail no longer repeat.
In one embodiment of the invention, the present invention can also provide a kind of processing unit, comprising: processor and storage Device, the memory are stored with computer program, and the processor runs the computer program to realize that preceding method is implemented Function in example or system embodiment.
For example, the memory, for storing the first computer program, the processor, for running described the One computer program is to realize medical history data processing method that such as Fig. 1 embodiment is shown.
For example, the memory, it can also be used to the doctor that storage is generated by the medical history data processing method Structured patient record data set is learned,
For example, the memory, it can also be used to store second computer program；The processor, for running Second computer program is stated to realize the function of the medical data recommender system shown such as Fig. 4 embodiment；For example, passing through operation Subprogram in second computer program realizes the process step of Fig. 5, Fig. 6 A (or 6B) embodiment or above-mentioned various other functions Suddenly.
The memory may include random access memory (RandomAccessMemory, abbreviation RAM), it is also possible to also Including nonvolatile memory (non-volatilememory), for example, at least a magnetic disk storage.
The processor can be general processor, including central processing unit (CentralProcessingUnit, letter Claim CPU), network processing unit (NetworkProcessor, abbreviation NP) etc.；It can also be digital signal processor (DigitalSignalProcessing, abbreviation DSP), specific integrated circuit (ApplicationSpecificIntegrated Circuit, abbreviation ASIC), field programmable gate array (Field-ProgrammableGateArray, abbreviation FPGA) or Other programmable logic device, discrete gate or transistor logic, discrete hardware components etc..
In one embodiment of the invention, computer storage medium can also be provided, store the first computer program, described the One computer program is used to run by processor to realize the medical history data processing method shown such as Fig. 1 embodiment.
In one embodiment of the invention, computer storage medium can also be provided, store by the medical history data The medical structure medical record data collection that processing method generates.
In one embodiment of the invention, computer storage medium can also be provided, store second computer program, described the Two computer programs are used to run by processor to realize the medical data recommender system shown such as Fig. 4 embodiment.
Above-mentioned computer readable storage medium can be any usable medium or include one that computer can access The data storage devices such as a or multiple usable mediums integrated server, data center.The usable medium can be magnetic Jie Matter, (for example, floppy disk, hard disk, tape), optical medium (for example, DVD) or semiconductor medium (such as solid state hard disk (SolidStateDisk, SSD)) etc..
Of the invention ensure that doctor can promote the validity of interrogation relative to traditional interrogation mode in terms of two.
Firstly, a large amount of medical record data can be converted to based on current computer information level and computing capability Structured patient record data set, the ability of the remote superman's class doctor of the medical record data amount that computer can be handled are analyzed by data The degree of association between the symptom information of discovery and between symptom information and symptom relevant information is also truer and objective.
Secondly, medical data recommender system of the invention can carry out symptom letter based on above structure medical record data collection Correlation analysis between the information such as breath, symptom relevant information is high to predict and recommend to meet the interrogation foundation in efficient interrogation direction Effect auxiliary Chinese medicinal interrogation process, the realistic meaning with height.
The above-described embodiments merely illustrate the principles and effects of the present invention, and is not intended to limit the present invention.It is any ripe The personage for knowing this technology all without departing from the spirit and scope of the present invention, carries out modifications and changes to above-described embodiment.Cause This, includes that institute is complete without departing from the spirit and technical ideas disclosed in the present invention for usual skill in technical field such as At all equivalent modifications or change, should be covered by the claims of the present invention.

Claims (21)

1. a kind of medical history data processing method characterized by comprising
It by each case history that a medical history is concentrated and its include that the content of each project is expressed as by mathematical character mode Project information under one structured patient record data and its various classifications of the items for being included, to be formed comprising described in one or more The medical structure medical record data collection of structured patient record data.
2. medical history data processing method according to claim 1, which is characterized in that the classification of the items includes: Symptomatic classification and the relative symptom correlation comprising one or more symptom relevant informations containing one or more symptom informations point Class；The described method includes:
Various unduplicated symptom informations are extracted from the medicine structure data set, storage forms symptom information library；
Each symptom relevant classification is corresponded to from the medicine structure data set extracts various unduplicated symptom correlation letters respectively Breath, storage form relevant information library corresponding with each symptom relevant classification.
3. medical history data processing method according to claim 2, which is characterized in that the medicine belongs to Chinese medicine neck Domain, the symptom relevant classification include: in patient basis, sign, medical history, card type and prescription any one or Multiple combinations.
4. medical history data processing method according to claim 2 characterized by comprising extract symptom information library In each symptom information between, and/or each symptom information and relevant information library in degree of association data between each symptom relevant information, and It is stored.
5. medical history data processing method according to claim 4, which is characterized in that in the extraction symptom information library Various symptoms information between degree of association data, comprising:
Every kind of symptom information in corresponding symptom information library, establishes symptom and vector occurs；Wherein, there is vector in the symptom It is expressed as the set for the characterization value whether every kind of symptom information appears in each structured patient record data；
It calculates symptom corresponding to various symptoms information and the similarity information between vector occurs, as the degree of association data.
6. medical history data processing method according to claim 5, which is characterized in that the calculating of the similarity information Method includes: in COS distance, Euclidean distance, standardization Euclidean distance, mahalanobis distance, Hamming distance and manhatton distance Any one.
7. medical history data processing method according to claim 4, which is characterized in that in the extraction symptom information library Each symptom information and relevant information library in degree of association data between each symptom relevant information, comprising:
Various symptoms in the case of each symptom relevant information occurs and symptom information library are counted in relevant information library respectively The quantity for the structured patient record data that information occurs accounts for the ratio of structured patient record data set respectively, with approximate representation for every kind of disease There is the conditional probability of every kind of symptom information in the case of occurring in shape relevant information, as the degree of association data.
8. according to claim 7, which is characterized in that the medicine belongs to traditional Chinese medical science field, and the relevant information library includes: Card type information bank comprising one or more card type information, and/or the prescription information library comprising one or more prescription informations.
9. medical history data processing method according to claim 8, which is characterized in that include master in the prescription information Prescription information；The main prescription information is for substituting the prescription information.
10. a kind of medical data recommender system, which is characterized in that the processing dress for being stored with structured patient record data set applied to one It sets；Wherein, the structured patient record data set includes at least one structured patient record data, each structured patient record data And its project information under the various classifications of the items for being included is the case history concentrated to a medical history and its includes each The mathematical character of the content of a project；The classification of the items include: the Symptomatic classification comprising one or more symptom informations and and its The relevant symptom relevant classification comprising one or more symptom relevant informations；The processing unit is also stored with: gathering the knot The default symptom information library and difference that structure medical record data concentrates various unduplicated default symptom informations to be formed are various described Each phase that various unduplicated default symptom relevant informations are formed in the medicine structure data set that classification of the items is gathered Close information bank；The processing unit is also stored with: between each default symptom information in the default symptom information library, and/or each Degree of association data in default symptom information and relevant information library between each default symptom relevant information；Institute's system includes:
Receiving unit, for obtaining the symptom information sequence to be diagnosed comprising one group of symptom information to be diagnosed；
Processing unit, for matching and calculating in the degree of association data between the default symptom information each in symptom information library The degree of association data between each default symptom information, set form association symptom to one group of symptom information to be diagnosed respectively Sequence pushes away so that default symptom information corresponding to selection wherein one or more maximum data of magnitude of degree of association data is used as Recommend symptom information；
And/or
The processing unit is calculated for being inputted using the structured patient record data set as data based on each described wait diagnose The posterior probability that the default symptom relevant information in each of relevant information library that symptom information occurs occurs；In the respectively default symptom Select maximum one of posterior probability values as recommending symptom relevant information in relevant information；Alternatively, by posterior probability values maximum Multiple default symptom relevant informations alternately symptom relevant information, judge close between each alternative symptom relevant information Whether the magnitude of degree evidence is lower than preset value；If it is not, then selecting the maximum alternative symptom relevant information of posterior probability values as pushing away Recommend symptom relevant information；If so, each default symptom information and each default disease in relevant information library in the symptom information library In degree of association data between shape relevant information, each alternative symptom relevant information target between each default symptom information respectively is matched Degree of association data, and according to the target association degree being matched to according to calculating: it is relevant to all different from each symptom information to be diagnosed Each of diversity factor data between the alternative symptom relevant information of default symptom information, set form project and distinguish symptom sequence Column, for choosing default symptom information corresponding to wherein one or more maximum data of magnitude of diversity factor data as recommendation Symptom information is as recommendation symptom information.
11. medical data recommender system according to claim 10, which is characterized in that the pass between the default symptom information Connection degree evidence includes: the similarity letter that symptom corresponding to each default symptom information occurs between vector in default symptom information library Breath；Vector, which occurs, in the symptom to be established according to each default symptom information of correspondence, for indicating each default symptom Whether information appears in the set of the characterization value in each structured patient record data；
In degree of association data between the default symptom information each in symptom information library, matches and calculate described one group and wait for The degree of association data between each default symptom information, set form association symptom sequence to diagnostic symptom information respectively, comprising:
In degree of association data between default symptom information, matching is each described to be believed to diagnostic symptom with each default symptom Current degree of association data between breath；
Summed result is respectively obtained to each current degree of association data relevant to each default symptom information, with each summed result Set construct the association symptom sequence；
Wherein, in the association symptom sequence, the magnitude of the summed result is bigger, corresponding to default symptom information make To recommend the priority of symptom information higher.
12. medical data recommender system according to claim 11, which is characterized in that the similarity information is by remaining Any one calculating in chordal distance, Euclidean distance, standardization Euclidean distance, mahalanobis distance, Hamming distance and manhatton distance What mode obtained.
13. medical data recommender system according to claim 10, which is characterized in that each pre- in the symptom information library It is each in relevant information library if the degree of association data in symptom information and relevant information library between each default symptom relevant information Symptom relevant information occur in the case of and symptom information library in every kind of symptom information occur structured patient record data quantity The ratio of structured patient record data set is accounted for respectively, is every kind of disease occur in the case of every kind of symptom relevant information occurs with approximate representation The conditional probability of shape information, as the degree of association data.
14. medical data recommender system according to claim 10, which is characterized in that the multiple alternative symptom correlation letter Breath includes: the default symptom relevant information that posterior probability values are maximum and take second place；
The alternative symptom relevant information for being relevant to default symptom informations all each of different from each symptom information to be diagnosed Between diversity factor data, comprising: the posterior probability values are maximum and the default symptom relevant information taken second place correspond to it is each pre- If one group of target association degree of symptom information is not less than zero difference between.
15. medical data recommender system described in 3 or 14 according to claim 1, which is characterized in that the medicine belongs to Chinese medicine neck Domain, the relevant information library include: the card type information bank comprising one or more card type information, and/or comprising one or more prescriptions The prescription information library of information.
16. medical data recommender system according to claim 15, which is characterized in that include main side in the prescription information Agent information；The main prescription information is for substituting the prescription information.
17. medical data recommender system according to claim 11, which is characterized in that the processing unit, for pressing institute It distributes weight and chooses corresponding respectively power respectively from the association symptom sequence and one or more described project identification symptom sequences The maximum data of quantity magnitude of weight, and the corresponding default symptom information of those data is gathered to form comprehensive identification Symptom sequence.
18. medical data recommender system according to claim 17, which is characterized in that the processing unit is also used to root The weight is adjusted according to interrogation phase difference.
19. a kind of medical data recommended method, which is characterized in that the processing dress for being stored with structured patient record data set applied to one It sets；Wherein, the structured patient record data set includes at least one structured patient record data, each structured patient record data And its project information under the various classifications of the items for being included is the case history concentrated to a medical history and its includes each The mathematical character of the content of a project；The classification of the items include: the Symptomatic classification comprising one or more symptom informations and and its The relevant symptom relevant classification comprising one or more symptom relevant informations；The processing unit is also stored with: gathering the knot The default symptom information library and difference that structure medical record data concentrates various unduplicated default symptom informations to be formed are various described Each phase that various unduplicated default symptom relevant informations are formed in the medicine structure data set that classification of the items is gathered Close information bank；The processing unit is also stored with: between each default symptom information in the default symptom information library, and/or each Degree of association data in default symptom information and relevant information library between each default symptom relevant information；Institute's method includes:
Obtain the symptom information sequence to be diagnosed comprising one group of symptom information to be diagnosed；
In degree of association data in symptom information library between each default symptom information, matches and calculate described one group and wait diagnosing The degree of association data between each default symptom information, set form association symptom sequence to symptom information respectively, for choosing Wherein default symptom information corresponding to one or more maximum data of the magnitude of degree of association data is as recommendation symptom information；
And/or
It is inputted using the structured patient record data set as data, calculates the correlation occurred based on each symptom information to be diagnosed Each of information bank presets the posterior probability that symptom relevant information occurs；
Select in the respectively default symptom relevant information posterior probability values maximum one as recommendation symptom relevant information；Or Person judges the maximum multiple default symptom relevant informations of posterior probability values alternately symptom relevant information each alternative Whether the magnitude of the proximity data between symptom relevant information is lower than preset value；If it is not, then selecting posterior probability values maximum standby Select symptom relevant information as recommendation symptom relevant information；If so, in the symptom information library each default symptom information with In degree of association data in relevant information library between each default symptom relevant information, each alternative symptom relevant information is matched respectively and respectively Target association degree evidence between default symptom information, and according to the target association degree being matched to according to calculating: be relevant to it is each Diversity factor data between the alternative symptom relevant information of all each of different default symptom informations of symptom information to be diagnosed, set Formation project distinguishes symptom sequence, pre- corresponding to wherein one or more maximum data of magnitude of diversity factor data for choosing If symptom information is as recommendation symptom information as recommendation symptom information.
20. a kind of processing unit, which is characterized in that including processor and memory；
The memory, for storing the first computer program,
The processor, for running first computer program to realize doctor as claimed in any one of claims 1-9 wherein Learn medical record data processing method；
Alternatively, the memory, for storing the medical history data processing method as described in any one of claims 1 to 9 The medical structure medical record data collection of generation；
Alternatively, the memory, for storing second computer program；
The processor, for running the second computer program to realize as described in any one of claim 10 to 18 Medical data recommender system realizes medical data recommended method as claimed in claim 19.
21. a kind of computer storage medium, which is characterized in that the first computer program of storage, first computer program are used In being run by processor to realize medical history data processing method as claimed in any one of claims 1-9 wherein；Alternatively, depositing Store up the medical structure medical record data collection that the medical history data processing method as described in any one of claims 1 to 9 generates； Alternatively, storage second computer program, the second computer program is used to run by processor to realize such as claim 10 To medical data recommender system described in any one of 18.
CN201810902948.4A 2018-08-09 2018-08-09 Medical history data processing method, medical data recommender system, equipment and medium Pending CN109119134A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810902948.4A CN109119134A (en) 2018-08-09 2018-08-09 Medical history data processing method, medical data recommender system, equipment and medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810902948.4A CN109119134A (en) 2018-08-09 2018-08-09 Medical history data processing method, medical data recommender system, equipment and medium

Publications (1)

Publication Number Publication Date
CN109119134A true CN109119134A (en) 2019-01-01

Family

ID=64851965

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810902948.4A Pending CN109119134A (en) 2018-08-09 2018-08-09 Medical history data processing method, medical data recommender system, equipment and medium

Country Status (1)

Country Link
CN (1) CN109119134A (en)

Cited By (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109817329A (en) * 2019-01-21 2019-05-28 暗物智能科技(广州)有限公司 A kind of medical treatment interrogation conversational system and the intensified learning method applied to the system
CN109887562A (en) * 2019-02-20 2019-06-14 广州天鹏计算机科技有限公司 The similarity of electronic health record determines method, apparatus, equipment and storage medium
CN109935337A (en) * 2019-02-25 2019-06-25 长沙学院 A kind of medical record lookup method and system based on similarity measurement
CN109949885A (en) * 2019-03-12 2019-06-28 重庆医事通科技发展有限公司 A kind of tele-medicine file data optimizing polymerization system and optimum management method
CN110706772A (en) * 2019-10-11 2020-01-17 北京百度网讯科技有限公司 Sorting method and device, electronic equipment and storage medium
CN111161881A (en) * 2019-06-13 2020-05-15 南京医基云医疗数据研究院有限公司 Method and device for identifying disease co-occurrence relationship and storage medium
CN112397194A (en) * 2019-08-16 2021-02-23 北京大数医达科技有限公司 Method, device and electronic equipment for generating patient condition attribution interpretation model
WO2021139116A1 (en) * 2020-05-14 2021-07-15 平安科技（深圳）有限公司 Method, apparatus and device for intelligently grouping similar patients, and storage medium
CN113268511A (en) * 2021-04-21 2021-08-17 广东易生活信息科技有限公司 Ancient book and ancient prescription based traditional Chinese medicine prescription recommendation method and system
CN113327655A (en) * 2021-04-21 2021-08-31 福建亿能达信息技术股份有限公司 Outlier detection method, device, equipment and medium for multidimensional data
CN113658710A (en) * 2021-08-11 2021-11-16 东软集团股份有限公司 Data matching method and related equipment thereof
CN113707277A (en) * 2021-08-04 2021-11-26 浙江卡易智慧医疗科技有限公司 Prostate MR cancer structured report design method
CN113707276A (en) * 2021-08-04 2021-11-26 浙江卡易智慧医疗科技有限公司 Left and right carotid arteries and branch structured report design method thereof

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2004288047A (en) * 2003-03-24 2004-10-14 Fujitsu Ltd Medical examination support system and medical examination support program
CN102722633A (en) * 2011-03-28 2012-10-10 佳能株式会社 Medical diagnosis support apparatus and medical diagnosis support method
CN107145744A (en) * 2017-05-08 2017-09-08 合肥工业大学 Construction method, device and the aided diagnosis method of medical knowledge collection of illustrative plates
CN107330289A (en) * 2017-07-10 2017-11-07 叮当（深圳）健康机器人科技有限公司 A kind of symptom information analysis method and device

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2004288047A (en) * 2003-03-24 2004-10-14 Fujitsu Ltd Medical examination support system and medical examination support program
CN102722633A (en) * 2011-03-28 2012-10-10 佳能株式会社 Medical diagnosis support apparatus and medical diagnosis support method
CN107145744A (en) * 2017-05-08 2017-09-08 合肥工业大学 Construction method, device and the aided diagnosis method of medical knowledge collection of illustrative plates
CN107330289A (en) * 2017-07-10 2017-11-07 叮当（深圳）健康机器人科技有限公司 A kind of symptom information analysis method and device

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
孙艳秋 等: "中医电子病历中科研数据的结构化研究", 《信息系统工程》 *
李中平 等: "相似度分析方法在亚健康状态症状指标分析中的应用", 《中国卫生统计》 *
瞿年清 等: "中医智能(辅助)诊疗系统的设计", 《朱文锋学术思想研讨会暨中医诊断师资班30周年纪念大会论文集》 *

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109817329A (en) * 2019-01-21 2019-05-28 暗物智能科技(广州)有限公司 A kind of medical treatment interrogation conversational system and the intensified learning method applied to the system
CN109887562A (en) * 2019-02-20 2019-06-14 广州天鹏计算机科技有限公司 The similarity of electronic health record determines method, apparatus, equipment and storage medium
CN109935337A (en) * 2019-02-25 2019-06-25 长沙学院 A kind of medical record lookup method and system based on similarity measurement
CN109935337B (en) * 2019-02-25 2021-01-15 长沙学院 Medical record searching method and system based on similarity measurement
CN109949885A (en) * 2019-03-12 2019-06-28 重庆医事通科技发展有限公司 A kind of tele-medicine file data optimizing polymerization system and optimum management method
CN111161881A (en) * 2019-06-13 2020-05-15 南京医基云医疗数据研究院有限公司 Method and device for identifying disease co-occurrence relationship and storage medium
CN112397194B (en) * 2019-08-16 2024-02-06 北京大数医达科技有限公司 Method, device and electronic equipment for generating patient disease attribution interpretation model
CN112397194A (en) * 2019-08-16 2021-02-23 北京大数医达科技有限公司 Method, device and electronic equipment for generating patient condition attribution interpretation model
CN110706772A (en) * 2019-10-11 2020-01-17 北京百度网讯科技有限公司 Sorting method and device, electronic equipment and storage medium
WO2021139116A1 (en) * 2020-05-14 2021-07-15 平安科技（深圳）有限公司 Method, apparatus and device for intelligently grouping similar patients, and storage medium
CN113327655A (en) * 2021-04-21 2021-08-31 福建亿能达信息技术股份有限公司 Outlier detection method, device, equipment and medium for multidimensional data
CN113327655B (en) * 2021-04-21 2022-08-05 福建亿能达信息技术股份有限公司 Outlier detection method, device, equipment and medium for multidimensional data
CN113268511B (en) * 2021-04-21 2023-08-25 广东易生活信息科技有限公司 Ancient book-based traditional Chinese medicine prescription recommendation method and system
CN113268511A (en) * 2021-04-21 2021-08-17 广东易生活信息科技有限公司 Ancient book and ancient prescription based traditional Chinese medicine prescription recommendation method and system
CN113707277A (en) * 2021-08-04 2021-11-26 浙江卡易智慧医疗科技有限公司 Prostate MR cancer structured report design method
CN113707276A (en) * 2021-08-04 2021-11-26 浙江卡易智慧医疗科技有限公司 Left and right carotid arteries and branch structured report design method thereof
CN113658710A (en) * 2021-08-11 2021-11-16 东软集团股份有限公司 Data matching method and related equipment thereof

Similar Documents

Publication Publication Date Title
CN109119134A (en) Medical history data processing method, medical data recommender system, equipment and medium
Abdar et al. NE-nu-SVC: a new nested ensemble clinical decision support system for effective diagnosis of coronary artery disease
CN111191020B (en) Prescription recommendation method and system based on machine learning and knowledge graph
Yoon et al. Discovery and clinical decision support for personalized healthcare
Sanabria et al. Remote patient monitoring program in automated peritoneal dialysis: impact on hospitalizations
Miled et al. Predicting dementia with routine care EMR data
Chattopadhyay et al. A Case‐Based Reasoning system for complex medical diagnosis
Rashid et al. An augmented artificial intelligence approach for chronic diseases prediction
Marateb et al. Manipulating measurement scales in medical statistical analysis and data mining: A review of methodologies
Lu et al. Collaborative graph learning with auxiliary text for temporal event prediction in healthcare
CN106650292B (en) A kind of individual health record system that function is supported with process decision
CN109785927A (en) Clinical document structuring processing method based on internet integration medical platform
US20130231957A1 (en) Intelligent filtering of health-related information
Alonso et al. Evaluation of semantic similarity metrics applied to the automatic retrieval of medical documents: An UMLS approach
CN103049641A (en) Medical guidance diagnosis display method and system
Lu et al. Understanding heart failure patients EHR clinical features via SHAP interpretation of tree-based machine learning model predictions
US20220020457A1 (en) Medical data evaluation utilization system and medical data evaluation utilization method
CN109213871A (en) Patient information knowledge mapping construction method, readable storage medium storing program for executing and terminal
Kadi et al. Systematic mapping study of data mining–based empirical studies in cardiology
Huang et al. Knowledge-based patient data generation
Kunjir et al. Big data analytics and visualization for hospital recommendation using HCAHPS standardized patient survey
CN105701330A (en) Health information processing method and system
Afsar et al. An exploration on-demand article recommender system for cancer patients information provisioning
Stiglic et al. Pediatric readmission classification using stacked regularized logistic regression models
Song et al. Developing high risk clusters for chronic disease events with classification association rule mining

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20190101

RJ01 Rejection of invention patent application after publication