CN109636421A - Medical data exception recognition methods, equipment and storage medium based on machine learning - Google Patents
Medical data exception recognition methods, equipment and storage medium based on machine learning Download PDFInfo
- Publication number
- CN109636421A CN109636421A CN201811530957.1A CN201811530957A CN109636421A CN 109636421 A CN109636421 A CN 109636421A CN 201811530957 A CN201811530957 A CN 201811530957A CN 109636421 A CN109636421 A CN 109636421A
- Authority
- CN
- China
- Prior art keywords
- medical data
- data
- chinese herbal
- herbal medicine
- medical
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000010801 machine learning Methods 0.000 title claims abstract description 32
- 238000003860 storage Methods 0.000 title claims abstract description 18
- 241000411851 herbal medicine Species 0.000 claims abstract description 137
- 230000002159 abnormal effect Effects 0.000 claims abstract description 46
- 238000012549 training Methods 0.000 claims abstract description 25
- 230000002547 anomalous effect Effects 0.000 claims abstract description 18
- 238000012216 screening Methods 0.000 claims abstract description 17
- 230000033228 biological regulation Effects 0.000 claims abstract description 14
- 239000003814 drug Substances 0.000 claims description 42
- 229940079593 drug Drugs 0.000 claims description 26
- 238000012545 processing Methods 0.000 claims description 12
- 238000002790 cross-validation Methods 0.000 claims description 7
- 238000012795 verification Methods 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 5
- 230000000306 recurrent effect Effects 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 4
- 238000012423 maintenance Methods 0.000 claims description 4
- 230000006399 behavior Effects 0.000 claims description 2
- 235000013399 edible fruits Nutrition 0.000 claims description 2
- 238000004458 analytical method Methods 0.000 abstract description 6
- 238000013473 artificial intelligence Methods 0.000 abstract description 3
- 238000005516 engineering process Methods 0.000 abstract description 3
- 230000008569 process Effects 0.000 description 11
- 238000010586 diagram Methods 0.000 description 10
- 230000007246 mechanism Effects 0.000 description 9
- 238000003745 diagnosis Methods 0.000 description 7
- 230000036541 health Effects 0.000 description 7
- 208000024891 symptom Diseases 0.000 description 7
- 238000004422 calculation algorithm Methods 0.000 description 6
- 244000025254 Cannabis sativa Species 0.000 description 5
- 201000010099 disease Diseases 0.000 description 5
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 5
- 238000000554 physical therapy Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 241000208340 Araliaceae Species 0.000 description 3
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 3
- 235000003140 Panax quinquefolius Nutrition 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 235000008434 ginseng Nutrition 0.000 description 3
- 210000000056 organ Anatomy 0.000 description 3
- 206010011224 Cough Diseases 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 239000006185 dispersion Substances 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 229910052602 gypsum Inorganic materials 0.000 description 2
- 239000010440 gypsum Substances 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 206010003658 Atrial Fibrillation Diseases 0.000 description 1
- 208000010392 Bone Fractures Diseases 0.000 description 1
- 208000003174 Brain Neoplasms Diseases 0.000 description 1
- FGUUSXIOTUKUDN-IBGZPJMESA-N C1(=CC=CC=C1)N1C2=C(NC([C@H](C1)NC=1OC(=NN=1)C1=CC=CC=C1)=O)C=CC=C2 Chemical compound C1(=CC=CC=C1)N1C2=C(NC([C@H](C1)NC=1OC(=NN=1)C1=CC=CC=C1)=O)C=CC=C2 FGUUSXIOTUKUDN-IBGZPJMESA-N 0.000 description 1
- 241000190633 Cordyceps Species 0.000 description 1
- 206010017076 Fracture Diseases 0.000 description 1
- 208000007882 Gastritis Diseases 0.000 description 1
- 206010028980 Neoplasm Diseases 0.000 description 1
- 241001248610 Ophiocordyceps sinensis Species 0.000 description 1
- 206010034464 Periarthritis Diseases 0.000 description 1
- 206010037660 Pyrexia Diseases 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 238000001467 acupuncture Methods 0.000 description 1
- 230000001154 acute effect Effects 0.000 description 1
- 230000001754 anti-pyretic effect Effects 0.000 description 1
- 239000002221 antipyretic Substances 0.000 description 1
- 239000010839 body fluid Substances 0.000 description 1
- 230000001684 chronic effect Effects 0.000 description 1
- 208000023652 chronic gastritis Diseases 0.000 description 1
- 206010009887 colitis Diseases 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 206010012601 diabetes mellitus Diseases 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 230000006806 disease prevention Effects 0.000 description 1
- 208000000718 duodenal ulcer Diseases 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 230000002107 myocardial effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000001717 pathogenic effect Effects 0.000 description 1
- 230000007170 pathology Effects 0.000 description 1
- 208000008494 pericarditis Diseases 0.000 description 1
- 230000035479 physiological effects, processes and functions Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000004224 protection Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 210000004243 sweat Anatomy 0.000 description 1
- 239000006188 syrup Substances 0.000 description 1
- 235000020357 syrup Nutrition 0.000 description 1
- 238000002560 therapeutic procedure Methods 0.000 description 1
- 208000029761 vertebral disease Diseases 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/018—Certifying business or products
- G06Q30/0185—Product, service or business identity fraud
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/08—Insurance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Tourism & Hospitality (AREA)
- Technology Law (AREA)
- Entrepreneurship & Innovation (AREA)
- Educational Administration (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Primary Health Care (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The medical data exception recognition methods based on machine learning that the invention discloses a kind of, it include: to receive that medical institutions acquires and the history of insured people that uploads is gone to a doctor data, and go to a doctor based on default screening rule from history and filter out the normal medical data for containing Chinese herbal medicine and the medical data of exception as the medical data for meeting modeling in data；Chinese herbal medicine expense Exception Model is established using machine learning mode as training sample using the medical data filtered out；Medical data to be detected are obtained, and medical data to be detected are directed into Chinese herbal medicine expense Exception Model and carry out anomalous identification, obtain recognition result；If the recognition result of medical data to be detected is abnormal, it is determined that medical data to be detected are that Chinese herbal medicine is claimed for charge at random in violation of rules and regulations.The invention also discloses a kind of equipment and storage mediums.The present invention realizes the intellectual analysis of data by the machine learning in artificial intelligence technology, to precisely efficiently identify out suspicious medical data, and effectively controls the medical care precess of violation.
Description
Technical field
The present invention relates to technical field of data processing more particularly to a kind of medical data exception identifications based on machine learning
Method, equipment and storage medium.
Background technique
Social medical insurance is state and society according to certain laws and regulations, is provided for the labourer into scope of insurance coverage
Basic medical demand guarantee when illness and the social security system established.Currently, medical service organ is widely distributed, it is insured people
Member provides more convenience-for-people medical services, however under the driving of interests, some mechanisms occur making unwarranted inspections, disorderly write a prescription, repeat just
It examines repetition the violation operations such as write a prescription, collect fees arbitrarily, wastes originally limited medical resource, while also compromising insurant
Legitimate interests, and then it is unfavorable for the sustainable development of quality of medical care and medical treatment & health.
Summary of the invention
The medical data exception recognition methods that the main purpose of the present invention is to provide a kind of based on machine learning, equipment and
Storage medium, it is intended to solve how precisely to efficiently identify out suspicious medical data, to effectively control the medical treatment behaviour of violation
The technical issues of making.
To achieve the above object, a kind of medical data exception recognition methods based on machine learning provided by the invention, institute
State medical data exception recognition methods the following steps are included:
Receive that medical institutions acquires and the history of insured people that uploads is gone to a doctor data, and based on presetting screening rule from described
History, which is gone to a doctor, filters out the normal medical data for containing Chinese herbal medicine and abnormal medical data in data as meeting going to a doctor for modeling
Data；
Chinese herbal medicine expense exception mould is established using machine learning mode as training sample using the medical data filtered out
Type；
Medical data to be detected are obtained, and the medical data to be detected are directed into the Chinese herbal medicine expense Exception Model
Anomalous identification is carried out, recognition result is obtained；
If the recognition result of the medical data to be detected is abnormal, it is determined that the medical data to be detected are Chinese herbal medicine
It claims for charge at random in violation of rules and regulations.
Optionally, described to be filtered out from the medical data of the history containing the normal of Chinese herbal medicine based on default screening rule
Medical data and abnormal medical data include: as the medical data for meeting modeling
According to the data content of the medical data of the history, the medical data of the history without containing Chinese herbal medicine are rejected, wherein institute
Data content is stated including at least consultation time, medical medication and drug cost；
According to consultation time, Chinese herbal medicine and its drug cost in the medical data of remaining history, statistics is obtained in single
Herbal medicine total cost and single month Chinese herbal medicine total cost；
Whether the single Chinese herbal medicine total cost and/or single month Chinese herbal medicine total cost for judging the medical data of history are greater than or equal to
Preset threshold；
If history is gone to a doctor, the single Chinese herbal medicine total cost of data and/or single month Chinese herbal medicine total cost are greater than or equal to default threshold
History data of going to a doctor then are labeled as abnormal medical data by value；
If history is gone to a doctor, the single Chinese herbal medicine total cost of data and/or single month Chinese herbal medicine total cost are less than preset threshold,
History data of going to a doctor are labeled as data of normally going to a doctor；
Using the medical data of the exception of mark and the normally medical data as the medical data for meeting modeling.
Optionally, described that Chinese herbal medicine is established using machine learning mode as training sample using the medical data filtered out
Expense Exception Model includes:
Processing is extracted to the medical data filtered out, obtains characteristic parameter and timing information；
It is gone to a doctor using the history of insured people corresponding to the medical data that filter out data of going to a doctor as input quantity, with this data
Obtained characteristic parameter and timing information are handled as output quantity, using recurrent neural network to the input quantity and the output
The value sample of amount is trained, and obtains Chinese herbal medicine expense Exception Model.
Optionally, the described pair of medical data filtered out extract processing, obtain characteristic parameter and timing information includes:
Data cleansing, participle slice are successively carried out to the medical data filtered out and remove unrelated word, obtains entry, and
Feature extraction is carried out to the entry, obtains characteristic parameter；
According to time location of the characteristic parameter in medical data, the corresponding timing information of characteristic parameter is determined.
Optionally, after the operation for obtaining Chinese herbal medicine expense Exception Model, the medical data exception identification side
Method further include:
Cross validation is rolled over using K or the Chinese herbal medicine expense Exception Model is verified in random intersection Holdout verifying,
It is verified as a result, so that maintenance personnel safeguards or changes to the Chinese herbal medicine expense Exception Model according to the verification result
Into.
Optionally, described that the medical data to be detected are directed into the abnormal knowledge of Chinese herbal medicine expense Exception Model progress
Not, obtaining recognition result includes:
The medical data to be detected are directed into the Chinese herbal medicine expense Exception Model and carry out anomalous identification, to obtain
State the index value of medical data to be detected；
Judge whether the index value is greater than or equal to preset standard value；
If the index value is greater than or equal to preset standard value, it is determined that the recognition result of the medical data to be detected is
It is abnormal；
If the index value is less than preset standard value, it is determined that the recognition result of the medical data to be detected is normal.
Optionally, medical data to be detected are obtained described, and the medical data to be detected is directed into the medium-height grass
Expenses for medicine carries out anomalous identification with Exception Model, and after obtaining the operation of recognition result, the medical data exception recognition methods is also
Include:
The medical data to be detected are merged with the medical data filtered out, to update the Chinese herbal medicine expense
The training sample of Exception Model；
According to the training sample of update, the Chinese herbal medicine expense Exception Model is further trained, with
To the Chinese herbal medicine expense Exception Model of update.
Optionally, if the recognition result in the medical data to be detected is abnormal, it is determined that it is described it is to be detected just
Examining data is the medical data exception recognition methods after the operation that Chinese herbal medicine is claimed for charge at random in violation of rules and regulations further include:
Obtain the corresponding insured people's information of the abnormal medical data to be detected and medical institutions' information；
According to insured people's information, right-safeguarding notice is issued to insured people；
According to medical institutions' information, the notice that notifies and/or deduct fees in violation of rules and regulations is issued to medical institutions.
In addition, to achieve the above object, the present invention also provides a kind of medical data exceptions to identify equipment, the medical data
Anomalous identification equipment includes: memory, processor and is stored in the number that can be run on the memory and on the processor
According to anomalous identification program, realized as described in any one of above-mentioned when the data exception recognizer is executed by the processor
The step of medical data exception recognition methods based on machine learning.
In addition, to achieve the above object, it is described computer-readable the present invention also provides a kind of computer readable storage medium
It is stored with data exception recognizer on storage medium, realizes when the data exception recognizer is executed by processor as above-mentioned
Any one of described in the medical data exception recognition methods based on machine learning the step of.
The present invention passes through first receives that medical institutions acquires and the history of insured people that uploads is gone to a doctor data, and based on presetting
Screening rule is gone to a doctor from history filters out the normal medical data for containing Chinese herbal medicine and abnormal medical data in data as meeting
The medical data of modeling；Chinese herbal medicine expense is established using machine learning mode as training sample using the medical data filtered out
Exception Model.Then obtain medical data to be detected, and by medical data to be detected be directed into Chinese herbal medicine expense Exception Model into
Row anomalous identification obtains recognition result, and then identifies medical data with the presence or absence of abnormal conditions.If recognition result is exception,
Then determine that medical data to be detected are that Chinese herbal medicine is claimed for charge at random in violation of rules and regulations.The present invention is realized by the machine learning in artificial intelligence technology
The intellectual analysis of data to precisely efficiently identify out suspicious medical data, and effectively controls the medical care precess of violation, into
And the dynamics of supervision medical institutions is improved, Perfecting Supervision mechanism, and then promote the sustainable development of quality of medical care and medical treatment & health.
Detailed description of the invention
Fig. 1 is the structural schematic diagram that the medical data exception that the embodiment of the present invention is related to identifies equipment operating environment；
Fig. 2 is that the present invention is based on the flow diagrams of medical one embodiment of data exception recognition methods of machine learning；
Fig. 3 is the refinement flow diagram of mono- embodiment of Fig. 2 step S20；
Fig. 4 is that medical data to be detected are directed into Chinese herbal medicine expense Exception Model in Fig. 2 step S30 to carry out abnormal knowledge
Not, the refinement flow diagram of one embodiment of recognition result is obtained；
Fig. 5 is that the present invention is based on the flow diagrams of medical another embodiment of data exception recognition methods of machine learning.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific embodiment
It should be appreciated that described herein, specific examples are only used to explain the present invention, is not intended to limit the present invention.
Referring to Fig.1, Fig. 1 is that the medical data exception that the embodiment of the present invention is related to identifies the structure of equipment operating environment
Schematic diagram.
The medical data exception identification equipment of the present embodiment may include such as mobile phone, tablet computer, laptop, palm
The computer installations such as computer, server.
As shown in Figure 1, the medical data exception identification equipment may include: processor 1001, such as CPU, communication bus
1002, user interface 1003, network interface 1004, memory 1005.Wherein, communication bus 1002 for realizing these components it
Between connection communication.User interface 1003 may include display screen (Display), input unit such as keyboard (Keyboard),
Network interface 1004 optionally may include standard wireline interface and wireless interface (such as WI-FI interface).Memory 1005 can be with
It is high speed RAM memory, is also possible to stable memory (non-volatile memory), such as magnetic disk storage.Storage
Device 1005 optionally can also be the storage device independently of aforementioned processor 1001.
It will be understood by those skilled in the art that the hardware configuration for data exception identification equipment of going to a doctor shown in Fig. 1 is not
The restriction to medical data exception identification equipment is constituted, may include components more more or fewer than diagram, or combine certain
Component or different component layouts.
As shown in Figure 1, as may include operating system, net in a kind of memory 1005 of computer readable storage medium
Network communication module, Subscriber Interface Module SIM and computer program.Wherein, operating system is to manage and control medical data exception to know
The program of other equipment and software resource supports the operation of data exception recognizer and other softwares and/or program.
In the hardware configuration of medical data exception identification equipment shown in Fig. 1, network interface 1004 is mainly used for accessing
Network；User interface 1003 is mainly used for detecting confirmation Command And Edit instruction etc..And processor 1001 can be used for calling and deposit
The data exception recognizer stored in reservoir 1005, and execute the medical data exception recognition methods below based on machine learning
Each embodiment in operation.
Device hardware structure is identified based on above-mentioned medical data exception, proposes that the present invention is based on the medical data of machine learning
Each embodiment of abnormality recognition method.
It is that the present invention is based on the processes of medical one embodiment of data exception recognition methods of machine learning to show referring to Fig. 2, Fig. 2
It is intended to.
The embodiment of the invention provides the embodiments of medical data exception recognition methods, it should be noted that although flowing
Logical order is shown in blit, but in some cases, it can be to be different from shown or described by sequence execution herein
The step of.
In the present embodiment, the medical data exception recognition methods based on machine learning includes:
Step S10, receives that medical institutions acquires and the history of insured people that uploads is gone to a doctor data, and is advised based on default screening
It then goes to a doctor from history and filters out the normal medical data for containing Chinese herbal medicine and abnormal medical data in data as meeting modeling
Medical data；
In the present embodiment, social medical insurance is state and society according to certain laws and regulations, for into scope of insurance coverage
The labourer guarantee of basic medical demand and social security system for establishing when illness is provided.And insured is exactly to participate in social insurance
Abbreviation, insured people namely refer to participate in social insurance insurant.Medical institutions refer to that sets up in a manner prescribed by law is engaged in
Medical diagnosis on disease, the general name for treating movable health organ, including hospital, sanatorium, health care institute, clinic, clinic, disease prevention and cure
It stands, Health Center, clinic and first-aid station etc., more convenience-for-people medical services is provided for insurant.
In the present embodiment, patient be also simultaneously insured people seen a doctor in medical institutions it is medical, will generate it is a large amount of it is relevant just
Data are examined, personal information (age, gender, date of birth, identification card number etc.), consultation time, physician office visits including insured people,
Visit symptom and its corresponding physical therapy project and physical therapy expense, corresponding detection project and testing cost, corresponding medical medication
With the contents such as drug cost.Insured people all medical letters on the day of medical institutions go to a doctor and pass through brush social security card and can display real-time
Breath.The medical data of history are that insured people is gone to a doctor from first time to medical institutions to backward medical produced in same mechanism every time
Data.Various regions medical institutions establish with this medical data exception identification equipment communicate to connect respectively, realize data interaction.Medical treatment
Mechanism acquires the medical data of history of all insured people and is aggregated into local data base, then by the medical number of all history of acquisition
Relevant treatment is carried out according to this medical data exception identification equipment is uploaded to, and then each insured people is in each medical institutions, difference
The information that period goes to a doctor can be very clear.It is understood that history is gone to a doctor, data are with insured artificial independent individual one
The complete data of part；And various regions medical institutions can independent operation, can data under independent acquisition oneself mechanism, and
Not by the premeditated intervention of other mechanisms or insured people etc..
In the present embodiment, data volume is quickly accumulated under big data environment, to analyze the valence that mass data is contained
Value, it is particularly significant to filter out valuable data.And data screening is in vitally in entire flow chart of data processing
Position.The increase of data volume and the complication of structure, this makes the data screening towards big data that will necessarily expend more resource, because
This selects suitable algorithm very necessary to quick and precisely filter out valuable data.And each algorithm suffers from oneself and makes
With required environment, the increase of data complexity improves the difficulty of selection appropriate algorithm under big data environment；Secondly big number
According to huge data volume but also to analyze valuable data by single algorithm more and more difficult.Thus the screening of data exists
The first step for occupying critically important status and mathematical modeling in mathematical modeling, when and the only data got well just can guarantee
The authenticity and accuracy of the result obtained.It is filtered out from the medical data of each history based on default screening rule and meets modeling
Medical data are exactly the screening in magnanimity initial data (the medical data of the history for multiple insured people that more medical institutions upload)
The medical data of representational history are provided, and then are prepared for data mining, mathematical modeling.Specific data screening process is under
It is described in detail in literary embodiment.
Step S20 establishes medium-height grass expenses for medicine using machine learning mode as training sample using the medical data filtered out
Use Exception Model；
In the present embodiment, model is the mathematical model constructed with mathematical logic method and mathematical linguistics, and machine learning
Computer is exactly allowed to learn knowledge new out from existing data, that is, according to the medical data filtered out as training number
According to the study of the system of progress, for example how to identify data exception etc..Briefly, developer only needs mass data that is, will
The medical data filtered out input to computer, then sum up Data analysis logic therein by computer oneself, summarize
Corresponding logical code, to obtain a Chinese herbal medicine expense Exception Model.Trained process is exactly to utilize training sample and tie
Sample label, sample characteristics corresponding to data are closed, i.e., existing data determine the process of model parameter.Chinese herbal medicine expense is abnormal
Model can be local disposition and be also possible to cloud to be disposed online, and preferably cloud is disposed online, can real-time update data, optimization
The accuracy and reliability of Chinese herbal medicine expense Exception Model, while distributed can also extend, further satisfaction is complicated and changeable
The practical medical situation of insured people.
It should be added that Chinese herbal medicine expense Exception Model is to receive Chinese herbal treatment for this kind of issued by doctor
Insured people medical data and the anomalous identification model established, and then more acurrate can reliably identify medical number to be detected
Chinese herbal medicine, excessive Chinese herbal treatment, this kind of abnormal medical situation of arbitrary imposition of fees are such as repeated out according to whether there is.
Step S30 obtains medical data to be detected, and medical data to be detected is directed into Chinese herbal medicine expense Exception Model
Anomalous identification is carried out, recognition result is obtained；
In the present embodiment, also insured people is to the medical generated data of medical institutions for medical data to be detected, only
Unlike one, which is not intended to model and be used to detect whether that there are abnormal conditions, and the data can be ginseng
The medical data or the medical data of nearly half a year of guarantor closely several times are also possible to the medical data of history, do not do any restriction herein.
Pass through Chinese herbal medicine expense Exception Model, so that it may quickly recognize medical data to be detected with the presence or absence of such as repeating out medium-height grass
Medicine, excessive Chinese herbal treatment, this kind of abnormal medical situation of arbitrary imposition of fees, to contain medical institutions' violation operation, and ensure
The legitimate interests of insurant.It should be understood that medical data to be detected are the data for including Chinese herbal medicine content, Jin Ertong
Existing exception can targetedly be identified by crossing Chinese herbal medicine expense Exception Model.
Step S40, if the recognition result of medical data to be detected is abnormal, it is determined that medical data to be detected are Chinese herbal medicine
It claims for charge at random in violation of rules and regulations.
In the present embodiment, the recognition result of medical data to be detected is exception, illustrates that insured people is medical by unreasonable
Chinese herbal treatment and charge, at the same time, medical institutions must violation operations.Chinese herbal medicine violation arbitrary imposition of fees i.e. insured people's quilt
Doctor, which issues, receives unreasonable Chinese herbal medicine expense.Further, for abnormal medical data, to the medical institutions of violation into
The corresponding punishment of row, such as the prescription power, the fine that reduce the credibility of medical institutions, cancel doctor, it is serious or even can revoke
The practicing requirements etc. of medical institutions, are configured with specific reference to actual conditions.It should be understood that if medical data to be detected
Recognition result be it is normal, then illustrate insured people it is medical by the reasonable treatment for closing rule, pay and rationally close the relevant administration expenses of rule
With there is no violation operations.
In the present embodiment, the medical data of the history for the insured people for acquiring and uploading by reception medical institutions first, and base
It goes to a doctor in default screening rule from history and filters out the normal medical data containing Chinese herbal medicine and abnormal medical data work in data
For the medical data for meeting modeling；Medium-height grass is established using machine learning mode as training sample using the medical data filtered out
Expenses for medicine Exception Model.Then medical data to be detected are obtained, and medical data to be detected are directed into Chinese herbal medicine expense exception
Model carries out anomalous identification, obtains recognition result, and then identifies medical data with the presence or absence of abnormal conditions.If recognition result is
It is abnormal, it is determined that medical data to be detected are that Chinese herbal medicine is claimed for charge at random in violation of rules and regulations.The present invention passes through the engineering in artificial intelligence technology
The intellectual analysis for realizing data is practised, to precisely efficiently identify out suspicious medical data, and effectively controls the medical treatment of violation
Operation, and then the dynamics of supervision medical institutions is improved, Perfecting Supervision mechanism, and then promote holding for quality of medical care and medical treatment & health
Supervention exhibition.
Further, based on the above embodiment, in the present embodiment, step S10 is based on presetting screening rule from each history just
It examines and filters out the normal medical data for containing Chinese herbal medicine and abnormal medical data in data as the medical data for meeting modeling, packet
It includes:
Step a goes to a doctor the data content of data according to history, rejects the history without containing Chinese herbal medicine and goes to a doctor data, wherein
Data content includes at least consultation time, medical medication and drug cost；
In the present embodiment, data of going to a doctor, personal information (age, gender, date of birth, identification card number including insured people
Deng), consultation time, physician office visits, visit symptom and its corresponding detection project and testing cost, corresponding physical therapy project and reason
The contents such as treatment expense, corresponding medical medication and drug cost.Visit symptom, such as atrial fibrillation, fracture, diabetes, Acute myocardial
Pericarditis, enterogastritis, brain tumor etc..Currently, the treatment type of medical institutions' setting can simply be divided into two classes, one kind is west
Doctor, one kind is Chinese medicine.Wherein, Chinese medicine (Traditional Chinese Medicine), refers generally to Chinese Han nationality's labour people
The medicine based on traditional medicine that the people create, so also referred to as Chinese medicine, is to study the diagnosis of Human Physiology, pathology and disease and prevent
The Men Xueke controlled etc..Human body is regarded as the entity of gas, shape, mind, is passed through by traditional Chinese medicine using yin-yang and five elements as theoretical basis
The method that ginseng is closed in " four methods of diagnosis " four methods of diagnosis seeks the cause of disease, characteristic of disease, sick position, analyzes the vital organs of the human body in the interpretation of the cause, onset and process of an illness and human body, channels and collaterals pass
It saves, the variation of qi-blood-body fluid, judge pathogenic and vital contention, and then obtain name of disease, conclude type of coming to testify, with diagnosis and treatment principle, formulate
" sweat, spit, under and temperature, it is clear, mend, disappear " etc. therapies, it is a variety of using Chinese medicine, acupuncture, massage, massage, cupping, the qigong, dietotherapy etc.
Treatment means make human body reach yin-yang and reconcile and rehabilitation.Medical medication can equally be divided into Western medicine and Chinese medicine, wherein Chinese medicine be with
The acquisition of Chinese traditional medicine theoretical direction processes, preparation, and illustration mechanism instructs the drug of clinical application, including cordyceps sinensis
Flower, Radix Angelicae Sinensis, astragali, gypsum etc..Simultaneously it is understood that different Chinese herbal medicines equally correspondence collects different drug costs, have
Body is arranged according to the actual situation.
In the present embodiment, due to the medical data of history for the insured people that each medical institutions upload, some includes Chinese herbal medicine
Related data, some do not have, and is the accuracy and reliability for improving model identification, to the history of original all insured people
Medical data are screened, and specifically, reject the medical data of the history without containing Chinese herbal medicine, and retain the history containing Chinese herbal medicine
Medical data, and then reduce operation, avoid data omission, avoid the training for influencing Chinese herbal medicine expense Exception Model.
Scheme to further understand the present invention, with the medical data instance of the history of insured people, the medical number of the history of insured people A
According to such as table 1, the medical data such as table 2 of the history of insured people B, the medical data such as table 3 of the history of insured people C；
| Consultation time | Visit symptom | Physical therapy project | Physical therapy expense | Medical medication | Drug cost |
| 2018-1-1 | Flu | Nothing | Nothing | Cold drug | 20 |
| 2018-1-6 | Flu | Nothing | Nothing | Cold drug | 20 |
| 2018-4-5 | Fever | Nothing | Nothing | Antipyretic | 45 |
| 2018-6-7 | Cough | Nothing | Nothing | Cough syrup | 15 |
Table 1
Table 2
Table 3
By table 1-3 it is found that the diagnosis records of insured people A do not have a Chinese herbal medicine, and during the diagnosis records of insured people B and C have
Herbal medicine, thus insured people A is rejected, retain the medical data of history of insured people B and C.
Step b, according to consultation time, Chinese herbal medicine and its drug cost in the medical data of remaining history, statistics is obtained
Single Chinese herbal medicine total cost and single month Chinese herbal medicine total cost；
In the present embodiment, further to help computer learning, the medical data of history after screening is classified, that is, are divided
For this abnormal and normal two major classes data, specifically, the consultation time in data of being gone to a doctor according to remaining each history, Chinese herbal medicine and
Its drug cost, statistics obtain single Chinese herbal medicine total cost and single month Chinese herbal medicine total cost.Due to the medical data of remaining history
There is Chinese herbal medicine, by taking insured people B and C as an example, statistics obtains insured people B respectively in 2018-4-12 single Chinese herbal medicine total cost
147, in 2018-4-17 single Chinese herbal medicine total cost 498, and then it is 645 that statistics, which obtains the single month Chinese herbal medicine total cost in April,；
Insured people C is in 2015-6-3 single/mono- month Chinese herbal medicine total cost 116.
Step c, judges whether the single Chinese herbal medicine total cost of the medical data of history and/or single month Chinese herbal medicine total cost are greater than
Or it is equal to preset threshold；
Step d, if history is gone to a doctor, the single Chinese herbal medicine total cost of data and/or single month Chinese herbal medicine total cost are greater than or equal to
History data of going to a doctor then are labeled as abnormal medical data by preset threshold；
Step e, if history is gone to a doctor, the single Chinese herbal medicine total cost of data and/or single month Chinese herbal medicine total cost are less than default threshold
History data of going to a doctor then are labeled as data of normally going to a doctor by value；
Step f, using the medical data of exception of mark and normal medical data as the medical data for meeting modeling.
In the present embodiment, since the quantity of the medical data of remaining history is big, with each insured artificial individual, gone through to each
The medical data of history are judged specifically, whether the single Chinese herbal medicine total cost for judging that statistics obtains is greater than or equal to default threshold
Whether value, single month Chinese herbal medicine total cost for judging that statistics obtains are greater than or equal to preset threshold, and then obtain the medical data of history
It is abnormal or normal.While to help computer learning, analysis is obtained into the abnormal or normal medical data of history and is stamped
Label, that is, do mark and distinguish, and then computer can data be classified as which kind by marking simple and clear knowing,
And then it is ready for training pattern.Further, threshold value, such as single medium-height grass are set for single/mono- month Chinese herbal medicine total cost
Medicine total cost threshold value is 400, and single month Chinese herbal medicine total cost threshold value is 1200 etc., is configured with specific reference to actual needs, in turn
The history that single Chinese herbal medicine total cost is more than 400 is gone to a doctor into data markers as abnormal data, single Chinese herbal medicine total cost is not surpassed
The medical data markers of history for crossing 400 are normal data, the medical data mark of the history for being more than 1200 for single month Chinese herbal medicine total cost
It is denoted as abnormal data, the history by single month Chinese herbal medicine total cost no more than 1200 goes to a doctor data markers as normal data.With insured
For people B and C, insured people B single expense is more than preset threshold, is labeled as abnormal medical data.Insured people C single/mono- monthly fee
With preset threshold is less than, it is labeled as data of normally going to a doctor.
In the present embodiment, either abnormal or normal data, as long as the history containing Chinese herbal medicine retained is medical
Data are to meet the medical data of modeling, these data be not be exactly normally abnormal, thus by the abnormal medical data of mark
Medical data with normal medical data as modeling.
It is the refinement flow diagram of mono- embodiment of Fig. 2 step S20 referring to Fig. 3, Fig. 3.
Based on the above embodiment, in the present embodiment, step S20, the medical data to filter out are used as training sample
Machine learning mode establishes Chinese herbal medicine expense Exception Model, comprising:
Step S21 extracts processing to the medical data filtered out, obtains characteristic parameter and timing information；
Specifically, step S21 includes:
1, data cleansing, participle slice are successively carried out to the medical data filtered out and remove unrelated word, obtain entry,
And feature extraction is carried out to entry, obtain characteristic parameter；
2, the time location according to characteristic parameter in medical data, determines the corresponding timing information of characteristic parameter.
In the present embodiment, although the medical data filtered out are the medical data of two kinds of different types (abnormal and normal),
But it requires to extract it processing, including data cleansing, participle are sliced and remove unrelated word: a. data cleansing, it is such as clear
Except the missing value of data, more dividing value, inconsistent code, repeated data etc..B. participle slice, as python stammerer participle, in
ICTCLAS tool of institute, section etc., can also condition random field (CRF) algorithm etc. based on statistics, network new word identification can pass through
New dictionary is manually added, or the solidified inside degree by calculating N-Grams is segmented come statistic frequency to design new word discovery and calculate
Method etc..C. unrelated word, such as removal punctuation mark, messy code, null, blank character are removed.If medical data are split into
The dry entry with meaning, entry includes that duodenal ulcer, chronic gastritis, chronic colitis, scapulohumeral periarthritis, lumbar vertebra disease etc. are this kind of
Symptom word；Including this kind of Chinese herbal medicine word such as cordyceps flower, Radix Angelicae Sinensis, astragali, gypsum；Including words such as expense, timing nodes.
In the present embodiment, characteristic parameter is the parameter information for characterizing substance or phenomenon characteristic, and feature extraction is
Entry is further processed, such as induction and conclusion, screening, obtained characteristic parameter include: visit symptom and its it is corresponding just
Examine medication, drug cost, consultation time, Chinese herbal medicine purchase number, medicine frequency；By data of going to a doctor, insurant is obtained just
Examine the expenditure that symptom always treats project for tumour, corresponding Chinese herbal medicine and its expenses for medicine, the frequency of continued treatment, Chinese herbal medicine expense Zhan
Than etc..
In the present embodiment, time series databases be mainly used for refer to processing band time tag (according to the time sequence variation,
I.e. the time serializes) data, the data with time tag are also referred to as time series data.The medical data in terms of time dimension,
Known to medical data refer to same target (insured people) acquired by different consultation times (changing according to the sequence of time)
Data.Chronologically-based feature must correspond to a time with regard to Chinese herbal medicine, the Chinese herbal medicine expense etc. of clinic purchase each time
The time location of point, i.e. characteristic parameter in medical data, and then determine the corresponding timing information of characteristic parameter.
Step S22, using the history of insured people corresponding to the medical data that filter out go to a doctor data as input quantity, with this
The characteristic parameter and timing information that medical data processing obtains are as output quantity, using recurrent neural network to input quantity and output
The value sample of amount is trained, and obtains Chinese herbal medicine expense Exception Model.
In the present embodiment, by two distinct types of medical data as the basic of training Chinese herbal medicine expense Exception Model
Data can make computer deep learning, and to cope with the medical data of different users, and then differentiation medical data to be detected are
It is no to belong to Exception Type.Filter out that meet the medical data of modeling be to be labeled as abnormal history and go to a doctor data and to be labeled as normal
History go to a doctor data, each history is gone to a doctor the corresponding insured people of data.Each medical data is extracted to obtain feature ginseng
Several and timing information.
In the present embodiment, recurrent neural network (Recurrent Neural Network, abbreviation RNN) be by addition across
The more connection hidden layer certainly at time point, and modeled in chronological order.Modeling is exactly to filter out the medical number for meeting modeling
It goes to a doctor as input quantity, with this characteristic parameter and timing that data processing obtains according to the history of corresponding insured people data of going to a doctor
Information uses following algorithm as output quantity, and in modeling process: 1. are randomly chosen k object, each object initially generation
The table center of one cluster；2. it is assigned to nearest cluster according to it at a distance from each cluster center by pair remaining each object；
3. recalculating the average value of each cluster, it is updated to new cluster center；4. 2,3 are constantly repeated, until criterion function convergence, training
Obtain Chinese herbal medicine expense Exception Model.It should be understood that the medical data of each insured people can regard a point as in a model,
By model, the dispersion degree of each insured people can be analyzed.
Further, it after obtaining Chinese herbal medicine expense Exception Model, also needs to verify the effect of model, specifically
Ground rolls over cross validation using K or Chinese herbal medicine expense Exception Model is verified in Holdout verifying, is verified as a result, in turn
Verify the recognition effect for the Chinese herbal medicine expense Exception Model trained.K rolls over cross validation, and initial samples are divided into K subsample,
One individual subsample is kept as the data of verifying model, other K-1 sample is used to train.Cross validation repeats K
Secondary, each subsample verifying is primary, and average K result or the other combinations of use finally obtain a single estimation.
Holdout verifying is not a kind of cross validation, because data there is no cross-reference, are to select portion from initial sample at random
Point, cross-validation data is formed, and it is remaining just as training data.In general, less than the number of script sample one third
According to being chosen as verify data.By the way that verification result is fed back to related maintenance personnel, and then maintains easily personnel and tied according to verifying
Fruit is safeguarded or is improved to Chinese herbal medicine expense Exception Model.For example the overall accuracy of verification result is not up to preset threshold such as
When 96%, maintenance personnel can analyze data according to the input quantity and output quantity of Chinese herbal medicine expense Exception Model, optimize operation method,
And then improve Chinese herbal medicine expense Exception Model；For example the overall accuracy of verification result can put into production when reaching preset threshold
It uses.
Referring to Fig. 4, Fig. 4 is that medical data to be detected are directed into Chinese herbal medicine expense Exception Model in Fig. 2 step S30 to carry out
Anomalous identification obtains the refinement flow diagram of one embodiment of recognition result.
Based on the above embodiment, in the present embodiment, it is different to be directed into Chinese herbal medicine expense by step S30 for medical data to be detected
Norm type carries out anomalous identification, obtains recognition result, comprising:
Step S31, by medical data to be detected be directed into Chinese herbal medicine expense Exception Model carry out anomalous identification, with obtain to
Detect the index value of medical data；
Whether step S32, judge index value are greater than or equal to preset standard value；
Step S33, if index value is greater than or equal to preset standard value, it is determined that the recognition result of medical data to be detected is
It is abnormal；
Step S34, if index value is less than preset standard value, it is determined that the recognition result of medical data to be detected is normal.
In the present embodiment, identification obtains index value, and index value can be consistent with the prior art herein, has very poor finger as used
Mark, quartile spacing index, eight quartile spacing indexs, equal poor index, standard deviation requirement etc., are set with specific reference to actual conditions
It sets.The bigger expression variation of index value is big, it is wide to spread, and the small expression deviation of value is small, comparatively dense.If medical data to be detected identify to obtain
Index value it is big, it is abnormal for illustrating dispersion degree greatly.If the index value that medical data to be detected identify is small, illustrate discrete
Small degree is normal.To determine whether medical data to be detected are abnormal, preset a standard value, preset standard convenient for machine
Value is the critical value of a normal medical data after the completion of model training, is determined as exception more than this critical value.
It is further alternative, in an alternative embodiment of the invention, after step S30, data exception recognition methods of going to a doctor
Further include:
Step A merges medical data to be detected with the medical data filtered out, different to update Chinese herbal medicine expense
The training sample of norm type；
Step B trains Chinese herbal medicine expense Exception Model, further according to the training sample of update to obtain
The Chinese herbal medicine expense Exception Model of update.
In the present embodiment, medical data to be detected are merged with the medical data filtered out, is using constantly newly-increased
Training sample, Chinese herbal medicine expense Exception Model is constantly adjusted and is optimized, the analysis data of medical data is increased, has
When helping analyze medical data to be detected, more typical and high accuracy analysis foundations are improved, and then improves exception and knows
Other accuracy.It should be clear that medical data to be detected have been detected and finished, by the way that the medical data detected are added, realize
Real-time update data, while updated training sample, it is more to be used for trained data, can advanced optimize and improve
Herbal medicine expense Exception Model.
It is that the present invention is based on the processes of medical another embodiment of data exception recognition methods of machine learning referring to Fig. 5, Fig. 5
Schematic diagram.
Based on the above embodiment, after the step s 40, medical data exception recognition methods further include:
Step S50 obtains the corresponding insured people's information of abnormal medical data to be detected and medical institutions' information；
Step S60 issues right-safeguarding notice to insured people according to insured people's information；
Step S70 issues the notice that notifies and/or deduct fees in violation of rules and regulations to medical institutions according to medical institutions' information.
In the present embodiment, medical data to be detected be insured people to medical institutions it is medical caused by data, including it is insured
People's information, while it being uploaded to server by medical institutions, the information such as the identity of medical institutions are carried, when to be detected medical
When data analysis is determined as abnormal, insured people and medical institutions can be determined by insured people's information, medical institutions' information, in turn
Correspondence is issued a notice.Right-safeguarding notice is issued to insured people, for informing user by unreasonable Chinese herbal treatment, informing user such as
What safeguards the lawful right of oneself in accordance with the law.Violation and/or notice of deducting fees are issued to medical institutions, for informing that medical institutions exist
Violation operation is imposed a fine it or is alerted, and then contains violations of rules and regulations, and improves the dynamics of supervision medical institutions, improves prison
Mechanism is superintended and directed, and then promotes the sustainable development of quality of medical care and medical treatment & health.Right-safeguarding notifies, notifies in violation of rules and regulations, notice of deducting fees it is interior
Appearance can write according to the actual situation, not do any restriction herein.
It should be added that the sequencing of step S60 and step S70 can be step S70 after first step S60,
It is also possible to step S60 after first step S70, is also possible to step S60 and step S70 while carrying out, do not do any restriction herein.
In addition, being stored on the computer readable storage medium the present invention also provides a kind of computer readable storage medium
There is data exception recognizer, is realized as described in any one of above-mentioned when the data exception recognizer is executed by processor
The step of medical data exception recognition methods based on machine learning.
Computer readable storage medium specific embodiment of the present invention and the above-mentioned medical data exception based on machine learning are known
Each embodiment of other method is essentially identical, and in this not go into detail.
It should be noted that, in this document, the terms "include", "comprise" or its any other variant are intended to non-row
His property includes, so that the process, method, article or the device that include a series of elements not only include those elements, and
And further include other elements that are not explicitly listed, or further include for this process, method, article or device institute it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including being somebody's turn to do
There is also other identical elements in the process, method of element, article or device.
The serial number of the above embodiments of the invention is only for description, does not represent the advantages or disadvantages of the embodiments.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side
Method can be realized by means of software and necessary general hardware platform, naturally it is also possible to by hardware, but in many cases
The former is more preferably embodiment.Based on this understanding, technical solution of the present invention substantially in other words does the prior art
The part contributed out can be embodied in the form of software products, which is stored in a readable storage medium
In matter (such as ROM/RAM, magnetic disk, CD), including some instructions are used so that a terminal (can be mobile phone, computer, service
Device, air conditioner or network equipment etc.) method that executes each embodiment of the present invention.
The embodiment of the present invention is described with above attached drawing, but the invention is not limited to above-mentioned specific
Embodiment, the above mentioned embodiment is only schematical, rather than restrictive, those skilled in the art
Under the inspiration of the present invention, without breaking away from the scope protected by the purposes and claims of the present invention, it can also make very much
Form, it is all using equivalent structure or equivalent flow shift made by description of the invention and accompanying drawing content, directly or indirectly
Other related technical areas are used in, all of these belong to the protection of the present invention.
The above is only a preferred embodiment of the present invention, is not intended to limit the scope of the invention, all to utilize this hair
Equivalent structure or equivalent flow shift made by bright specification and accompanying drawing content is applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of medical data exception recognition methods based on machine learning, which is characterized in that the medical data exception identification
Method the following steps are included:
Receive that medical institutions acquires and the history of insured people that uploads is gone to a doctor data, and based on presetting screening rule from the history
The normal medical data for containing Chinese herbal medicine and abnormal medical data are filtered out in medical data as the medical data for meeting modeling；
Chinese herbal medicine expense Exception Model is established using machine learning mode as training sample using the medical data filtered out；
Medical data to be detected are obtained, and the medical data to be detected are directed into the Chinese herbal medicine expense Exception Model and are carried out
Anomalous identification obtains recognition result；
If the recognition result of the medical data to be detected is abnormal, it is determined that the medical data to be detected are Chinese herbal medicine violation
Arbitrary imposition of fees.
2. medical data exception recognition methods as described in claim 1, which is characterized in that it is described based on default screening rule from
The history, which is gone to a doctor, filters out the normal medical data for containing Chinese herbal medicine and extremely medical data in data as meeting modeling
Medical data include:
According to the data content of the medical data of the history, the medical data of the history without containing Chinese herbal medicine are rejected, wherein the number
Consultation time, medical medication and drug cost are included at least according to content；
According to consultation time, Chinese herbal medicine and its drug cost in the medical data of remaining history, statistics obtains single Chinese herbal medicine
Total cost and single month Chinese herbal medicine total cost；
It is default to judge whether the single Chinese herbal medicine total cost of the medical data of history and/or single month Chinese herbal medicine total cost are greater than or equal to
Threshold value；
If history is gone to a doctor, the single Chinese herbal medicine total cost of data and/or single month Chinese herbal medicine total cost are greater than or equal to preset threshold,
History data of going to a doctor then are labeled as abnormal medical data；
If history is gone to a doctor, the single Chinese herbal medicine total cost of data and/or single month Chinese herbal medicine total cost are less than preset threshold, will go through
The medical data of history are labeled as data of normally going to a doctor；
Using the medical data of the exception of mark and the normally medical data as the medical data for meeting modeling.
3. medical data exception recognition methods as described in claim 1, which is characterized in that the medical data to filter out
As training sample, using machine learning mode, establishing Chinese herbal medicine expense Exception Model includes:
Processing is extracted to the medical data filtered out, obtains characteristic parameter and timing information；
It is gone to a doctor using the history of insured people corresponding to the medical data that filter out data of going to a doctor as input quantity, with this data processing
Obtained characteristic parameter and timing information is as output quantity, using recurrent neural network to the input quantity and the output quantity
Value sample is trained, and obtains Chinese herbal medicine expense Exception Model.
4. medical data exception recognition methods as claimed in claim 3, which is characterized in that the described pair of medical data filtered out
Processing is extracted, characteristic parameter is obtained and timing information includes:
Data cleansing, participle slice are successively carried out to the medical data filtered out and remove unrelated word, obtains entry, and to institute
Predicate item carries out feature extraction, obtains characteristic parameter；
According to time location of the characteristic parameter in medical data, the corresponding timing information of characteristic parameter is determined.
5. medical data exception recognition methods as claimed in claim 3, which is characterized in that obtain Chinese herbal medicine expense different described
After the operation of norm type, the medical data exception recognition methods further include:
Cross validation is rolled over using K or the Chinese herbal medicine expense Exception Model is verified in random intersection Holdout verifying, is obtained
Verification result, so that maintenance personnel safeguards or improves to the Chinese herbal medicine expense Exception Model according to the verification result.
6. medical data exception recognition methods as described in claim 1, which is characterized in that described by the medical number to be detected
Anomalous identification is carried out according to the Chinese herbal medicine expense Exception Model is directed into, obtaining recognition result includes:
The medical data to be detected are directed into the Chinese herbal medicine expense Exception Model and carry out anomalous identification, with obtain it is described to
Detect the index value of medical data；
Judge whether the index value is greater than or equal to preset standard value；
If the index value is greater than or equal to preset standard value, it is determined that the recognition result of the medical data to be detected is different
Often；
If the index value is less than preset standard value, it is determined that the recognition result of the medical data to be detected is normal.
7. medical data exception recognition methods as described in claim 1, which is characterized in that obtain medical number to be detected described
According to, and the medical data to be detected are directed into the Chinese herbal medicine expense Exception Model and carry out anomalous identification, obtain identification knot
After the operation of fruit, the medical data exception recognition methods further include:
The medical data to be detected are merged with the medical data filtered out, it is abnormal to update the Chinese herbal medicine expense
The training sample of model；
According to the training sample of update, the Chinese herbal medicine expense Exception Model is further trained, to obtain more
New Chinese herbal medicine expense Exception Model.
8. such as medical data exception recognition methods of any of claims 1-7, which is characterized in that if described
The recognition result of medical data to be detected is abnormal, it is determined that the medical data to be detected are the behaviour that Chinese herbal medicine is claimed for charge at random in violation of rules and regulations
After work, the medical data exception recognition methods further include:
Obtain the corresponding insured people's information of the abnormal medical data to be detected and medical institutions' information；
According to insured people's information, right-safeguarding notice is issued to insured people；
According to medical institutions' information, the notice that notifies and/or deduct fees in violation of rules and regulations is issued to medical institutions.
9. a kind of medical data exception identifies equipment, which is characterized in that the medical data exception identification equipment includes: storage
Device, processor and it is stored in the data exception recognizer that can be run on the memory and on the processor, the number
It is realized when being executed according to anomalous identification program by the processor and is based on machine learning as described in any item of the claim 1 to 8
Medical data exception recognition methods the step of.
10. a kind of computer readable storage medium, which is characterized in that it is different to be stored with data on the computer readable storage medium
Normal recognizer is realized when the data exception recognizer is executed by processor as described in any item of the claim 1 to 8
The step of medical data exception recognition methods based on machine learning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811530957.1A CN109636421A (en) | 2018-12-13 | 2018-12-13 | Medical data exception recognition methods, equipment and storage medium based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811530957.1A CN109636421A (en) | 2018-12-13 | 2018-12-13 | Medical data exception recognition methods, equipment and storage medium based on machine learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109636421A true CN109636421A (en) | 2019-04-16 |
Family
ID=66073939
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811530957.1A Pending CN109636421A (en) | 2018-12-13 | 2018-12-13 | Medical data exception recognition methods, equipment and storage medium based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109636421A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110222087A (en) * | 2019-05-15 | 2019-09-10 | 平安科技（深圳）有限公司 | Feature extracting method, device and computer readable storage medium |
| CN110910976A (en) * | 2019-10-12 | 2020-03-24 | 平安国际智慧城市科技股份有限公司 | Medical record detection method, device, equipment and storage medium |
| CN113869387A (en) * | 2021-09-18 | 2021-12-31 | 平安科技（深圳）有限公司 | Abnormal medical insurance reimbursement identification method and system based on artificial intelligence technology |
| CN114418008A (en) * | 2022-01-21 | 2022-04-29 | 平安国际智慧城市科技股份有限公司 | Medical treatment behavior identification method and device, terminal equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102013084A (en) * | 2010-12-14 | 2011-04-13 | 江苏大学 | System and method for detecting fraudulent transactions in medical insurance outpatient services |
| CN108197845A (en) * | 2018-02-28 | 2018-06-22 | 四川新网银行股份有限公司 | A kind of monitoring method of the transaction Indexes Abnormality based on deep learning model LSTM |
| CN108492196A (en) * | 2018-03-08 | 2018-09-04 | 平安医疗健康管理股份有限公司 | The air control method of medical insurance unlawful practice is inferred by data analysis |
| CN108597609A (en) * | 2018-05-04 | 2018-09-28 | 华东师范大学 | A kind of doctor based on LSTM networks is foster to combine health monitor method |
| CN108806780A (en) * | 2018-06-14 | 2018-11-13 | 四川久远银海软件股份有限公司 | A kind of exception medical expense judgment method and device |
| CN108921710A (en) * | 2018-06-08 | 2018-11-30 | 东莞迪赛软件技术有限公司 | The method and system of medical insurance abnormality detection |
-
2018
- 2018-12-13 CN CN201811530957.1A patent/CN109636421A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102013084A (en) * | 2010-12-14 | 2011-04-13 | 江苏大学 | System and method for detecting fraudulent transactions in medical insurance outpatient services |
| CN108197845A (en) * | 2018-02-28 | 2018-06-22 | 四川新网银行股份有限公司 | A kind of monitoring method of the transaction Indexes Abnormality based on deep learning model LSTM |
| CN108492196A (en) * | 2018-03-08 | 2018-09-04 | 平安医疗健康管理股份有限公司 | The air control method of medical insurance unlawful practice is inferred by data analysis |
| CN108597609A (en) * | 2018-05-04 | 2018-09-28 | 华东师范大学 | A kind of doctor based on LSTM networks is foster to combine health monitor method |
| CN108921710A (en) * | 2018-06-08 | 2018-11-30 | 东莞迪赛软件技术有限公司 | The method and system of medical insurance abnormality detection |
| CN108806780A (en) * | 2018-06-14 | 2018-11-13 | 四川久远银海软件股份有限公司 | A kind of exception medical expense judgment method and device |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110222087A (en) * | 2019-05-15 | 2019-09-10 | 平安科技（深圳）有限公司 | Feature extracting method, device and computer readable storage medium |
| CN110222087B (en) * | 2019-05-15 | 2023-10-17 | 平安科技（深圳）有限公司 | Feature extraction method, device and computer readable storage medium |
| CN110910976A (en) * | 2019-10-12 | 2020-03-24 | 平安国际智慧城市科技股份有限公司 | Medical record detection method, device, equipment and storage medium |
| WO2021068601A1 (en) * | 2019-10-12 | 2021-04-15 | 平安国际智慧城市科技股份有限公司 | Medical record detection method and apparatus, device and storage medium |
| CN113869387A (en) * | 2021-09-18 | 2021-12-31 | 平安科技（深圳）有限公司 | Abnormal medical insurance reimbursement identification method and system based on artificial intelligence technology |
| CN113869387B (en) * | 2021-09-18 | 2024-09-06 | 平安科技（深圳）有限公司 | Abnormal medical insurance reimbursement identification method and system based on artificial intelligence technology |
| CN114418008A (en) * | 2022-01-21 | 2022-04-29 | 平安国际智慧城市科技股份有限公司 | Medical treatment behavior identification method and device, terminal equipment and storage medium |
| CN114418008B (en) * | 2022-01-21 | 2024-09-13 | 深圳平安智慧医健科技有限公司 | Medical action recognition method, device, terminal equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109659035A (en) | Medical data exception recognition methods, equipment and storage medium based on machine learning | |
| CN109615012A (en) | Medical data exception recognition methods, equipment and storage medium based on machine learning | |
| CN109636421A (en) | Medical data exception recognition methods, equipment and storage medium based on machine learning | |
| Połap et al. | Agent architecture of an intelligent medical system based on federated learning and blockchain technology | |
| Tortorella et al. | Impacts of Healthcare 4.0 digital technologies on the resilience of hospitals | |
| CN107657536B (en) | The recognition methods of social security fraud and device | |
| CN109256212A (en) | Bone health assessment models construction method, device, equipment, medium and appraisal procedure | |
| CN109685671A (en) | Medical data exception recognition methods, equipment and storage medium based on machine learning | |
| Tanner-Smith et al. | Can brief alcohol interventions for youth also address concurrent illicit drug use? Results from a meta-analysis | |
| CN107066814A (en) | A kind of traditional Chinese medical science intelligent auxiliary diagnosis system cooperateed with based on the four methods of diagnosis | |
| CN105528526A (en) | Traditional Chinese medicine inheritance and big data mining-based dynamic health management system for life cycle | |
| Chu et al. | The use of artificial intelligence in complementary and alternative medicine: A systematic scoping review | |
| Cocos et al. | Crowd control: Effectively utilizing unscreened crowd workers for biomedical data annotation | |
| CN106933985B (en) | Analysis and discovery method of core party | |
| CN109616174A (en) | Recognition methods, device, terminal and the readable storage medium storing program for executing of repeated drug taking | |
| Chen et al. | Syndrome differentiation and treatment algorithm model in traditional Chinese medicine based on disease cause, location, characteristics and conditions | |
| KR20110098286A (en) | Self health diagnosis system of oriental medicine using fuzzy inference method | |
| Perri et al. | Multiple sensitive estimation and optimal sample size allocation in the item sum technique | |
| DE202023101305U1 (en) | An intelligent health and fitness data management system using artificial intelligence with IoT devices | |
| Warren et al. | Tightly bound: the relationship of network clustering coefficients and reincarceration at three therapeutic communities | |
| Huang et al. | Patterns exploration on patterns of empirical herbal formula of Chinese medicine by association rules | |
| Lee et al. | Fertility behaviors in South Korea and their association with ultrasound prenatal sex screening | |
| Ninness et al. | Behavioral and physiological neural network analyses: A common pathway toward pattern recognition and prediction | |
| Ramteke et al. | A Critical Analysis of The Methods Used To Classify Medical Data | |
| Dipro et al. | A federated learning approach for detecting Parkinson’s disease through privacy preserving by blockchain |
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: 20190416 |
|
| RJ01 | Rejection of invention patent application after publication |