CN107491638A - A kind of ICU user's prognosis method and terminal device based on deep learning model - Google Patents
A kind of ICU user's prognosis method and terminal device based on deep learning model Download PDFInfo
- Publication number
- CN107491638A CN107491638A CN201710633561.9A CN201710633561A CN107491638A CN 107491638 A CN107491638 A CN 107491638A CN 201710633561 A CN201710633561 A CN 201710633561A CN 107491638 A CN107491638 A CN 107491638A
- Authority
- CN
- China
- Prior art keywords
- data
- monitoring data
- user
- monitoring
- icu
- 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
Landscapes
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The embodiments of the invention provide a kind of ICU user's prognosis method and terminal device based on deep learning model, wherein, methods described includes：The Monitoring Data of critical illness ward ICU user is obtained, the Monitoring Data is to characterize the data of physiological index of user；Input using the Monitoring Data as deep learning model, prediction index data are calculated, the deep learning model is to train what is obtained according to the Historical Monitoring data and history achievement data of ICU user, and the prediction index data are used for change of illness state and probability of death of the instruction user after preset period of time；Prompted for the prediction index data.Using the present invention, the change of illness state and probability of death of ICU user after deep learning model prediction for a period of time can be utilized, to help doctor to provide effective reference information, helps to give treatment in time.
Description
Technical field
The present invention relates to medical science and information intelligent field, more particularly to a kind of ICU user based on deep learning model
Prognosis method and terminal device.
Background technology
Critical illness ward (Intensive Care Unit, ICU) is with the professional development of medical treatment and nursing, Novel medical
The birth of equipment and the improvement of hospital administrative structure and occur it is a kind of integrate modernization medical treatment and nursing technology medical team
Knit mode of management.ICU puts together urgent patient, and optimal guarantee is given on human and material resources and technology, is desirably to obtain good
Rescued effect.Therefore general all more critical, the disease of patient after how predicting a period of time of patient's condition in ICU wards is moved in
End of love and mortality risk, and intervened before the state of an illness is difficult to retrieve, it is a challenging task.
Therefore a kind of reasonable, efficient prediction scheme need to be proposed.
The content of the invention
Technical problem to be solved of the embodiment of the present invention is, there is provided a kind of ICU user based on deep learning model
Prognosis method and terminal device, using the change of illness state of ICU user after deep learning model prediction for a period of time and dead
Probability is died, in time, reliably to help doctor to give treatment to user, lifts practicality.
In a first aspect, the embodiments of the invention provide a kind of ICU user's prognosis method based on deep learning model,
Methods described includes：
The Monitoring Data of critical illness ward ICU user is obtained, the Monitoring Data is to characterize the data of physiological index of user；
Input using the Monitoring Data as deep learning model, prediction index data, the depth is calculated
It is to train what is obtained according to the Historical Monitoring data and history achievement data of ICU user to practise model, and the prediction index data are used
In change of illness state and probability of death of the instruction user after preset period of time；
Prompted for the prediction index data.
In some possible embodiments, the Monitoring Data for obtaining critical illness ward ICU includes：
The primary monitoring data of ICU user is obtained, the primary monitoring data is handled, obtains middle monitoring number
According to；
The middle Monitoring Data is converted into the Monitoring Data for possessing preset format.
In some possible embodiments, the primary monitoring data includes dynamic monitoring data and static monitoring techniques data,
Described that the primary monitoring data is handled, obtaining middle Monitoring Data includes：
The primary monitoring data is classified, obtains the dynamic monitoring data and the static monitoring techniques data；
The dynamic monitoring data and the static monitoring techniques data are pre-processed respectively, obtain dynamic intermediate data and
Static intermediate data, the pretreatment include at least one in the following：Data deduplication processing, dealing of abnormal data, data lack
Processing is filled up in mistake；
The dynamic intermediate data and the static intermediate data are merged, obtain the middle Monitoring Data.
It is described to carry out prompting for the prediction index data and include in some possible embodiments：
When the prediction index data exceed predetermined threshold value, the prediction index data are prompted, and the prediction is referred to
Mark data are sent to the contact person that prestores.
In some possible embodiments, the deep learning model is the deep learning model of N layers, and N is more than 0 just
Integer, the deep learning pattern include any one of following：Long memory network LSTM models in short-term, gate repeat unit network
GRU models, Recognition with Recurrent Neural Network RNN models, recurrent neural network RNNs models.
Second aspect, the embodiments of the invention provide a kind of terminal device, the terminal device includes above-mentioned for performing
The functional unit of the method for first aspect.
The third aspect, the embodiments of the invention provide a kind of terminal device, including：Processor, memory, communication interface and
Bus；The processor, the memory are connected by the bus with the communication interface and complete mutual communication；Institute
State memory storage executable program code；The processor is by reading the executable program code stored in the memory
To run program corresponding with the executable program code, for performing a kind of ICU user based on deep learning model
Prognosis method；Wherein, methods described is the method described in any one of first aspect.
Fourth aspect, the invention provides a kind of computer-readable recording medium, the computer-readable recording medium is deposited
The program code performed by computing device is stored up.Described program code includes being used for the method for performing any one in the first aspect
Instruction.
5th aspect, the invention provides a kind of computer program product including instructing, when it runs on computers
When so that the method that computer performs any one of above-mentioned first aspect.
The embodiment of the present invention can obtain the Monitoring Data of critical illness ward ICU user, the monitoring number by terminal device
According to characterize the data of physiological index of user, then the input using the Monitoring Data as deep learning model, is calculated
Prediction index data, the deep learning model are to be obtained according to the training of the Historical Monitoring data and history achievement data of ICU user
, the prediction index data are used for change of illness state and probability of death of the instruction user after preset period of time, finally for institute
Prediction index data are stated to be prompted；Between data of physiological index and prediction index data that ICU user can so be utilized
Relevance, the change of illness state and probability of death of user after predicting a period of time, in time, reliably to prompt, aid in doctor's treatment,
Improve the practicality of Forecasting Methodology.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
There is the required accompanying drawing used in technology description to be briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, on the premise of not paying creative work, can be with
Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 is that a kind of flow of ICU user's prognosis method based on deep learning model of the embodiment of the present invention is shown
It is intended to；
Fig. 2 is a kind of LSTM model schematics of the embodiment of the present invention；
Fig. 3 is a kind of stream of ICU user's prognosis method based on deep learning model of another embodiment of the present invention
Journey schematic diagram；
Fig. 4 is a kind of structural representation of terminal device of the embodiment of the present invention；
Fig. 5 is a kind of structural representation of terminal device of another embodiment of the present invention.
Embodiment
In order that those skilled in the art more fully understand the present invention program, below in conjunction with the embodiment of the present invention
Accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is only
The embodiment of a part of the invention, rather than whole embodiments.Based on the embodiment in the present invention, ordinary skill people
The every other embodiment that member is obtained under the premise of creative work is not made, it should all belong to the model that the present invention protects
Enclose.
Term " first ", " second " and " the 3rd " in description and claims of this specification and above-mentioned accompanying drawing is (such as
Fruit presence) etc. be used for distinguish different objects, not for description particular order.In addition, term " comprising " and they are any
Deformation, it is intended that cover non-exclusive include.Such as contain the process of series of steps or unit, method, system, product
Or equipment the step of being not limited to list or unit, but the step of alternatively also include not listing or unit, or can
Selection of land is also included for the intrinsic other steps of these processes, method, product or equipment or unit.
Fig. 1 is referred to, is a kind of ICU user's prognosis method based on deep learning model of the embodiment of the present invention
Schematic flow sheet, the methods described of the embodiment of the present invention can be applied in such as smart mobile phone, tablet personal computer, intelligence is wearable sets
It is standby to wait in the terminal device with communications network functionality, it can specifically be realized by the processor of these terminal devices.The present invention is implemented
The methods described of example also comprises the following steps.
Step S102, terminal device obtains the Monitoring Data of critical illness ward ICU user, and the Monitoring Data is used to characterize
The data of physiological index at family.
In this application, the Monitoring Data can be critical illness ward (Intensive Care Unit, ICU) patient
The currently monitored data of user, when the currently monitored data can be the data that monitor at current time or be current
Quarter for the previous period in the range of the data that are monitored.The Monitoring Data is the data for characterizing user's physical signs,
The data include but is not limited to heart rate data, breath data, temperature data, pulse data, height data, weight data etc.
Deng of the invention not limit.
Step S104, input of the described terminal device using the Monitoring Data as deep learning model, is calculated pre-
Achievement data is surveyed, the deep learning model is to train to obtain according to the Historical Monitoring data and history achievement data of ICU user
, the prediction index data are used for change of illness state and probability of death of the instruction user after preset period of time.
The deep learning model is that training in advance is stored in the terminal device, is specifically described below.Institute
State Historical Monitoring data and the history achievement data be the ICU wards user artificially counted real history data.It is described to go through
History achievement data is used for the probability of death for characterizing ICU wards user.
Step S106, prompted for the prediction index data.
The embodiment of the present invention is described below.
First, it is specific as follows that three kinds of embodiments are there are in step S102：
In the first embodiment, the terminal device obtains the primary monitoring data of ICU ward users.Then, it is described
Terminal device carries out the quantization of same data magnitude to the primary monitoring data, and the primary monitoring data is converted into and possessed
The Monitoring Data of unified preset format.
The primary monitoring data can be that terminal that the terminal device is detected using sensor or described is set
For what is obtained by network from server side, present embodiment does not limit.
The preset format includes but is not limited to the forms such as the decimal system, octodenary.Preferably, the application will be described original
Monitoring Data is converted into the Monitoring Data of integer type.For example, using the quantizing rule of ten series magnitudes, same order of magnitude amount is controlled
Monitoring Data after change is between 0 to 10.
In second of embodiment, the terminal device obtains the primary monitoring data of ICU ward users.Then, it is described
Terminal device is handled the primary monitoring data, such as duplicate removal processing etc., obtains middle Monitoring Data.Further,
The terminal device carries out the quantization of same data magnitude to the middle Monitoring Data, and the middle Monitoring Data is converted into
Possesses the Monitoring Data of unified preset format.
The first foregoing implementation is can be found on the acquisition of the primary monitoring data and the quantization of middle Monitoring Data
Associated description in mode, is repeated no more here.
In the third embodiment, the terminal device obtains the primary monitoring data of ICU ward users.Wherein, it is described
Primary monitoring data may include dynamic monitoring data and static monitoring techniques data.The terminal device enters to the primary monitoring data
Row processing, obtaining middle Monitoring Data includes：The terminal device is classified the primary monitoring data, is obtained described dynamic
State Monitoring Data and the static monitoring techniques data；The terminal device is respectively to the dynamic monitoring data and the static monitoring techniques
Data are pre-processed, and obtain dynamic intermediate data and static intermediate data；The terminal device is by the dynamic intermediate data
Merged with the static intermediate data, obtain the middle Monitoring Data, the pretreatment include it is following at least one
：Data deduplication processing, dealing of abnormal data, shortage of data fill up processing.
The dynamic monitoring data is can be with user's data of physiological index of time dynamic, such as user for sign
Heart rate data, pulse data, breath data etc..The static monitoring techniques data are not change over time or consolidate for characterizing
User's data of physiological index that (such as 1 month) will not change in timing is long, such as the data such as the height of user, body weight.
Exemplarily such as, terminal device availability data analysis module is respectively according to every item number in the primary monitoring data
The primary monitoring data is divided into two classes according to each self-corresponding mark (such as title), is dynamic monitoring data and static state
Monitoring Data.Then, the terminal device can be gone using Dynamic Data Processing module to dynamic monitoring data progress data
Weight, remove abnormal data, even fill up missing data etc. processing, obtaining middle dynamic data.Correspondingly, the terminal device
The static monitoring techniques data are learnt and trained etc. with processing using static data processing module, such as utilizes one-hot
Coding encodes to static data, obtains static intermediate data.Further, the terminal device availability data splicing mould
The dynamic intermediate data and the static intermediate data are spliced, merged by block, are supervised so as to obtain among complete, accurate
Survey data.Finally, the middle Monitoring Data can be converted into by the terminal device using Feature Engineering module possesses default lattice
The Monitoring Data of formula.
Data prediction is described in detail exemplified by filling up missing data below.Due to the prison of the user of hospital's collection
It is not rule to survey data, for example heart rate data per minute can all may be collected, but the frequency acquisition of temperature data is simultaneously
It is not fixed.Due in the handling process of dynamic monitoring data, it is necessary to by Monitoring Data (certainly including temperature data) unite
One is processed into the form of equal length.Therefore need to fill up missing values, because body temperature and time there may be linear relationship, can examine
Consider and fill up missing values using the method for linear regression.
In another example in the case of one kind is possible, oximetry data also per minute will not all collect correspondingly data.If
If keeping data format unified, filling up for missing values can be carried out using enthesis nearby, the purpose so handled is to make
Oximetry data will not produce larger fluctuation, so as to not interfere with data distortion or exception, not influence the study of model.
Before step S104, the terminal device can obtain the Historical Monitoring data of one or more groups of ICU wards users
With history achievement data.The Historical Monitoring data and the history achievement data can be the true of the collection of a period of time scope
Real data.Alternatively, the Historical Monitoring data and the history achievement data can prestore in the terminal device,
But obtained by network from other-end equipment or server, the present invention does not limit.Then the terminal device can
The Historical Monitoring data and the history achievement data are trained and learned using the mathematical modeling relevant with time series
Practise, so as to obtain deep learning model.
Correspondingly in step S104, the terminal device can be using the Monitoring Data obtained in step S102 as institute
The input of deep learning model is stated, the Monitoring Data is calculated by the deep learning model, so as to be predicted
Achievement data.The prediction index data are used for change of illness state and probability of death of the instruction user after preset period of time.
In an alternative embodiment, the deep learning model can be the deep learning model of n-layer, and n is just whole more than 0
Number.The deep learning pattern includes any one of following：Long memory network in short-term (Long Short-Term Memory,
LSTM) model, gate repeat unit network G RU models, Recognition with Recurrent Neural Network (Recurrent neural Network, RNN)
Model, recurrent neural network (Recurrent Neural Networks, RNNs) model, BP neural network model or other with
Data model of time series association etc., the present invention does not limit.
For example, by taking three-layer neural network LSTM models as an example.Such as Fig. 2, LSTM models include 2 layers of LSTM units (figure
It is shown as LSTMcell) and one layer of full connection (Fully Connected, FC) neutral net member (being illustrated as FC).Wherein every layer
LSTMcell and FC quantity does not limit.The ICU wards user that mode input end can gather at input terminal equipment each moment
Monitoring Data, calculated by three layers of LSTM models, from prediction index data corresponding to the output of model output end.
By taking the t4 moment as an example, mode input end inputs the Monitoring Data of moment ICU ward users, first layer and the second layer
LTSMcell the Monitoring Data of input can be carried out between feature extraction and study, such as learning data linearly or nonlinearly
Relation etc..Because LSTMcell allows information selectively to be transmitted, therefore the Monitoring Data before the t4 moment is (i.e.
Historical Monitoring data) pass through and selectively transmit and will predict ICU by FC layers together with the Monitoring Data at t4 moment
Ward user for a period of time after probability of death, by FC to model output end export calculate prediction index data.
It is very important in ICU scenes, because existing conventional method e insufficient to the Monitoring Data before backtracking, but
It is that the deteriorations of patient's physiological conditions may just appear clue before a period of time.Therefore work as and determined using conventional method
When patient is among being critically ill, it is likely that it is late, the chance of doctor's PCI can be given few, using energy of the present invention
Problems of the prior art are enough avoided, sufficient intervention time is reserved to doctor and removes symptom management.
In step S106, the terminal device can prompt the prediction index data.Mode on prompting is included but not
It is limited to the prompting modes such as sound, captions, picture, suspension windows, vibration, the embodiment of the present invention does not limit.
In an alternative embodiment, it is described to carry out prompting for the prediction index data and include：In the prediction index number
During according to more than predetermined threshold value, the prediction index data are prompted, and the prediction index data are sent to the contact person that prestores.
Specifically, when the prediction index data exceed predetermined threshold value, i.e., the probability of death of model prediction exceedes default
During threshold value, the treatment etc. in time of alert notification related doctor can be sent.Alternatively, the terminal device can also be by the prediction index
Data are sent to the contact person that prestores, such as are sent to hospital or cloud server, expert's treatment system etc., to aid in curing
It is raw to complete treatment.
The predetermined threshold value can be that user side or the terminal equipment side are independently set, and the embodiment of the present invention does not limit
It is fixed.
The terminal device can include user equipment (User Equipment, UE), smart mobile phone (such as Android hands
Machine, IOS mobile phones etc.), PC, tablet personal computer, palm PC, mobile internet device (MID, Mobile Internet
Devices) or internet device, the embodiment of the present invention such as wearable intelligent equipment are not construed as limiting.
The embodiment of the present invention can obtain critical illness ward ICU Monitoring Data by terminal device, and the Monitoring Data is
The data of physiological index of user is characterized, then the input using the Monitoring Data as deep learning model, is calculated prediction
Achievement data, the deep learning model are to train what is obtained according to the Historical Monitoring data and history achievement data of ICU user,
The prediction index data are used for change of illness state and probability of death of the instruction user after preset period of time, finally for the prediction
Achievement data is prompted；The relevance between the data of physiological index of ICU user and prediction index data can be so utilized,
The change of illness state and probability of death of user after predicting a period of time, in time, reliably to prompt, aid in doctor's treatment, improve
The practicality of Forecasting Methodology.
Fig. 3 is referred to, is another ICU user's prognosis based on deep learning model provided in an embodiment of the present invention
Method.Method as described in Figure 3, including step is implemented as follows：
Step S302, terminal device obtains the primary monitoring data of ICU user.
Step S304, described terminal device is classified the primary monitoring data, obtains the dynamic monitoring data
With the static monitoring techniques data；
Step S306, described terminal device is located in advance to the dynamic monitoring data and the static monitoring techniques data respectively
Reason, obtains dynamic intermediate data and static intermediate data, the pretreatment includes at least one in the following：At data deduplication
Reason, dealing of abnormal data, shortage of data fill up processing.
Step S308, described terminal device is merged the dynamic intermediate data and the static intermediate data, is obtained
To the middle Monitoring Data.
The middle Monitoring Data is converted into the Monitoring Data for possessing preset format by step S310, described terminal device.
Step S312, input of the described terminal device using the Monitoring Data as deep learning model, is calculated pre-
Achievement data is surveyed, the deep learning model is to train to obtain according to the Historical Monitoring data and history achievement data of ICU user
, the prediction index data are used for change of illness state and probability of death of the instruction user after preset period of time.
In an alternative embodiment, the deep learning model is the deep learning model of N layers, and N is the positive integer more than 0,
The deep learning pattern includes any one of following：Long memory network LSTM models in short-term, gate repeat unit network G RU moulds
Type, Recognition with Recurrent Neural Network RNN models, recurrent neural network RNNs models.
Step S314, the prediction index data are prompted when the prediction index data exceed predetermined threshold value, and by institute
State prediction index data and be sent to the contact person that prestores.
, can be referring specifically to the correlation in embodiment described in Fig. 1 on the part that the embodiment of the present invention is not shown and does not describe
Description, is repeated no more here.
Refer to Fig. 4, be a kind of structural representation of terminal device of the embodiment of the present invention, the embodiment of the present invention it is described
Terminal device 400 includes：Acquiring unit 402, processing unit 404 and Tip element 406；Wherein：
The acquiring unit 402, for obtaining critical illness ward ICU Monitoring Data, the Monitoring Data is used to characterize
The data of physiological index at family；
The processing unit 404, for the input using the Monitoring Data as deep learning model, prediction is calculated
Achievement data, the deep learning model are to train what is obtained according to the Historical Monitoring data and history achievement data of ICU user,
The prediction index data are used for change of illness state and probability of death of the instruction user after preset period of time；
The Tip element 406, for being prompted for the prediction index data.
In some possible embodiments, the acquiring unit 402 is used for the primary monitoring data for obtaining ICU user, right
The primary monitoring data is handled, and obtains middle Monitoring Data；The acquiring unit 402 is additionally operable to will the middle monitoring
Data are converted into the Monitoring Data for possessing preset format.
In some possible embodiments, the primary monitoring data includes dynamic monitoring data and static monitoring techniques data,
The acquiring unit 402, for the primary monitoring data to be classified, obtain the dynamic monitoring data and
The static monitoring techniques data；
The acquiring unit 402, it is additionally operable to respectively carry out in advance the dynamic monitoring data and the static monitoring techniques data
Processing, obtains dynamic intermediate data and static intermediate data, the pretreatment includes at least one in the following：At data deduplication
Reason, dealing of abnormal data, shortage of data fill up processing；
The acquiring unit 402, it is additionally operable to be merged the dynamic intermediate data and the static intermediate data, obtains
To the middle Monitoring Data.
In some possible embodiments,
The Tip element 406, for when the prediction index data exceed predetermined threshold value, prompting the prediction index
Data, and the prediction index data are sent to the contact person that prestores.
The specific implementation for the unit being related in the embodiment of the present invention refers to Fig. 1 to Fig. 3 and corresponds to correlation in embodiment
The description of functional unit or implementation steps, will not be described here.
Fig. 5 is referred to, Fig. 5 is a kind of structural representation of terminal device disclosed in the embodiment of the present invention.The present embodiment
Terminal device 400 includes：At least one processor 601, communication interface 602, user interface 603 and memory 604, processor
601st, communication interface 602, user interface 603 can be connected with memory 604 by bus or other manner, the embodiment of the present invention
Exemplified by being connected by bus 605.Wherein,
Processor 601 can be general processor, such as central processing unit (Central Processing Unit,
CPU)。
Communication interface 602 can be wireline interface (such as Ethernet interface) or wave point (such as cellular network interface
Or use wireless lan interfaces), for being communicated with other-end or website.In the embodiment of the present invention, communication interface 602
Specifically for obtaining Monitoring Data.
The concretely contact panel, including touch-screen and touch screen of user interface 603, for detecting the behaviour on contact panel
Instruct, user interface 603 can also be physical button or mouse.User interface 603 can also be display screen, for defeated
Go out, display image or data.
Memory 604 can include volatile memory (Volatile Memory), such as random access memory
(Random Access Memory, RAM)；Memory can also include nonvolatile memory (Non-Volatile
), such as read-only storage (Read-Only Memory, ROM), flash memory (Flash Memory), hard disk Memory
(Hard Disk Drive, HDD) or solid state hard disc (Solid-State Drive, SSD)；Memory 604 can also include upper
State the combination of the memory of species.Memory 604 is used to store batch processing code, and processor 601 is used to call memory 604
The program code of middle storage, perform following operation：
Critical illness ward ICU Monitoring Data is obtained, the Monitoring Data is to characterize the data of physiological index of user；
Input using the Monitoring Data as deep learning model, prediction index data, the depth is calculated
It is to train what is obtained according to the Historical Monitoring data and history achievement data of ICU user to practise model, and the prediction index data are used
In change of illness state and probability of death of the instruction user after preset period of time；
Prompted for the prediction index data.
In some possible embodiments, the processor 601 is specifically used for：
The primary monitoring data of ICU user is obtained, the primary monitoring data is handled, obtains middle monitoring number
According to；
The middle Monitoring Data is converted into the Monitoring Data for possessing preset format.
In some possible embodiments, the primary monitoring data includes dynamic monitoring data and static monitoring techniques data,
The processor 601 is specifically used for：
The primary monitoring data is classified, obtains the dynamic monitoring data and the static monitoring techniques data；
The dynamic monitoring data and the static monitoring techniques data are pre-processed respectively, obtain dynamic intermediate data and
Static intermediate data, the pretreatment include at least one in the following：Data deduplication processing, dealing of abnormal data, data lack
Processing is filled up in mistake；
The dynamic intermediate data and the static intermediate data are merged, obtain the middle Monitoring Data.
In some possible embodiments, the processor 601 is specifically used for：
When the prediction index data exceed predetermined threshold value, the prediction index data are prompted, and the prediction is referred to
Mark data are sent to the contact person that prestores.
In some possible embodiments, the deep learning model is the deep learning model of N layers, and N is more than 0 just
Integer, the deep learning pattern include any one of following：Long memory network LSTM models in short-term, gate repeat unit network
GRU models, Recognition with Recurrent Neural Network RNN models, recurrent neural network RNNs models.
The embodiment of the present invention also provides a kind of computer-readable storage medium, wherein, the computer-readable storage medium can be stored with journey
Sequence, including the part or all of implementation step in the methods described embodiment described in the above method embodiment when program performs
Suddenly.
It should be noted that for foregoing each method embodiment, in order to be briefly described, therefore it is all expressed as a series of
Combination of actions, but those skilled in the art should know, the present invention is not limited by described sequence of movement because
According to the present invention, some steps can use other orders or carry out simultaneously.Secondly, those skilled in the art should also know
Know, embodiment described in this description belongs to preferred embodiment, and involved action and module are not necessarily of the invention
It is necessary.
In the above-described embodiments, the description to each embodiment all emphasizes particularly on different fields, and does not have the portion being described in detail in some embodiment
Point, it may refer to the associated description of other embodiment.
In several embodiments provided herein, it should be understood that disclosed device, can be by another way
Realize.For example, device embodiment described above is only schematical, such as the division of the unit, it is only one kind
Division of logic function, can there is an other dividing mode when actually realizing, such as multiple units or component can combine or can
To be integrated into another system, or some features can be ignored, or not perform.Another, shown or discussed is mutual
Coupling direct-coupling or communication connection can be by some interfaces, the INDIRECT COUPLING or communication connection of device or unit,
Can be electrical or other forms.
The unit illustrated as separating component can be or may not be physically separate, show as unit
The part shown can be or may not be physical location, you can with positioned at a place, or can also be distributed to multiple
On NE.Some or all of unit therein can be selected to realize the mesh of this embodiment scheme according to the actual needs
's.
In addition, each functional unit in various embodiments of the present invention can be integrated in a processing unit, also may be used
To be that unit is individually physically present, can also two or more units it is integrated in a unit.It is above-mentioned integrated
Unit can both be realized in the form of hardware, can also be realized in the form of SFU software functional unit.
If the integrated unit is realized in the form of SFU software functional unit and is used as independent production marketing or use
When, it can be stored in a computer read/write memory medium.Based on such understanding, technical scheme is substantially
The part to be contributed in other words to prior art or all or part of the technical scheme can be in the form of software products
Embody, the computer software product is stored in a storage medium, including some instructions are causing a computer
Equipment (can be personal computer, server or network equipment etc.) perform each embodiment methods described of the present invention whole or
Part steps.And foregoing storage medium includes：USB flash disk, read-only storage (ROM, Read-Only Memory), arbitrary access are deposited
Reservoir (RAM, Random Access Memory), mobile hard disk, magnetic disc or CD etc. are various can be with store program codes
Medium.
Described above, the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although with reference to before
Embodiment is stated the present invention is described in detail, it will be understood by those within the art that：It still can be to preceding
State the technical scheme described in each embodiment to modify, or equivalent substitution is carried out to which part technical characteristic；And these
Modification is replaced, and the essence of appropriate technical solution is departed from the scope of various embodiments of the present invention technical scheme.
Claims (10)
- A kind of 1. ICU user's prognosis method based on deep learning model, it is characterised in that including：The Monitoring Data of critical illness ward ICU user is obtained, the Monitoring Data is to characterize the data of physiological index of user；Input using the Monitoring Data as deep learning model, prediction index data, the deep learning mould is calculated Type is to train what is obtained according to the Historical Monitoring data and history achievement data of ICU user, and the prediction index data are used to refer to Show change of illness state and probability of death of the user after preset period of time；Prompted for the prediction index data.
- 2. according to the method for claim 1, it is characterised in that the Monitoring Data for obtaining critical illness ward ICU user Including：The primary monitoring data of ICU user is obtained, the primary monitoring data is handled, obtains middle Monitoring Data；The middle Monitoring Data is converted into the Monitoring Data for possessing preset format.
- 3. according to the method for claim 2, it is characterised in that the primary monitoring data includes dynamic monitoring data and quiet State Monitoring Data, described that the primary monitoring data is handled, obtaining middle Monitoring Data includes：The primary monitoring data is classified, obtains the dynamic monitoring data and the static monitoring techniques data；The dynamic monitoring data and the static monitoring techniques data are pre-processed respectively, obtain dynamic intermediate data and static state Intermediate data, the pretreatment include at least one in the following：Data deduplication processing, dealing of abnormal data, shortage of data are filled out Benefit processing；The dynamic intermediate data and the static intermediate data are merged, obtain the middle Monitoring Data.
- 4. according to the method for claim 1, it is characterised in that described to carry out prompting bag for the prediction index data Include：When the prediction index data exceed predetermined threshold value, the prediction index data are prompted, and by the prediction index number According to being sent to the contact person that prestores.
- 5. according to the method any one of claim 1-4, it is characterised in that the deep learning model is the depth of N layers Learning model is spent, N is the positive integer more than 0, and the deep learning pattern includes any one of following：Long memory network in short-term LSTM models, gate repeat unit network G RU models, Recognition with Recurrent Neural Network RNN models, recurrent neural network RNNs models.
- A kind of 6. terminal device, it is characterised in that including：Acquiring unit, for obtaining the Monitoring Data of critical illness ward ICU user, the Monitoring Data is to characterize the physiology of user Achievement data；Processing unit, for the input using the Monitoring Data as deep learning model, prediction index data, institute is calculated It is to train what is obtained according to the Historical Monitoring data and history achievement data of ICU user to state deep learning model, and the prediction refers to Mark data are used for change of illness state and probability of death of the instruction user after preset period of time；Tip element, for being prompted for the prediction index data.
- 7. terminal device according to claim 6, it is characterised in thatThe acquiring unit, for obtaining the primary monitoring data of ICU user, the primary monitoring data is handled, obtained To middle Monitoring Data；The acquiring unit, it is additionally operable to for the middle Monitoring Data to be converted into the Monitoring Data for possessing preset format.
- 8. terminal device according to claim 7, it is characterised in that the primary monitoring data includes dynamic monitoring data With static monitoring techniques data,The acquiring unit, for the primary monitoring data to be classified, obtain the dynamic monitoring data and described quiet State Monitoring Data；The acquiring unit, it is additionally operable to respectively pre-process the dynamic monitoring data and the static monitoring techniques data, obtains To dynamic intermediate data and static intermediate data, the pretreatment includes at least one in the following：Data deduplication processing, exception Processing is filled up in data processing, shortage of data；The acquiring unit, it is additionally operable to be merged the dynamic intermediate data and the static intermediate data, obtains described Middle Monitoring Data.
- 9. terminal device according to claim 6, it is characterised in thatThe Tip element, for when the prediction index data exceed predetermined threshold value, prompting the prediction index data, and The prediction index data are sent to the contact person that prestores.
- 10. a kind of computer-readable recording medium, the computer-readable recording medium storage has computer program, and its feature exists In realization such as any one of claim 1 to 5 methods described when the computer program is executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710633561.9A CN107491638A (en) | 2017-07-28 | 2017-07-28 | A kind of ICU user's prognosis method and terminal device based on deep learning model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710633561.9A CN107491638A (en) | 2017-07-28 | 2017-07-28 | A kind of ICU user's prognosis method and terminal device based on deep learning model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107491638A true CN107491638A (en) | 2017-12-19 |
Family
ID=60644988
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710633561.9A Pending CN107491638A (en) | 2017-07-28 | 2017-07-28 | A kind of ICU user's prognosis method and terminal device based on deep learning model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107491638A (en) |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108073497A (en) * | 2018-01-29 | 2018-05-25 | 上海洞识信息科技有限公司 | A kind of multi objective unusual fluctuation analysis method based on data center's data acquisition platform |
| CN108182974A (en) * | 2017-12-27 | 2018-06-19 | 深圳和而泰数据资源与云技术有限公司 | Illness appraisal procedure, terminal device and computer-readable medium |
| CN108198617A (en) * | 2017-12-27 | 2018-06-22 | 深圳和而泰数据资源与云技术有限公司 | Illness appraisal procedure, terminal device and computer-readable medium |
| 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 |
| CN108877924A (en) * | 2018-03-30 | 2018-11-23 | 杭州依图医疗技术有限公司 | A kind of asthma method of determining probability and device |
| CN109359785A (en) * | 2018-11-30 | 2019-02-19 | 安徽四创电子股份有限公司 | Forced quarantine drug rehabilitation releasing personnel based on big data relapse hypothesis analysis system |
| CN109684320A (en) * | 2018-12-25 | 2019-04-26 | 清华大学 | The method and apparatus of monitoring data on-line cleaning |
| CN109887599A (en) * | 2019-02-25 | 2019-06-14 | 中国计量大学 | A kind of tcm prescription curative effect deduction method neural network based |
| CN110473636A (en) * | 2019-08-22 | 2019-11-19 | 山东众阳健康科技集团有限公司 | Intelligent doctor's advice recommended method and system based on deep learning |
| CN110911011A (en) * | 2019-11-27 | 2020-03-24 | 医惠科技有限公司 | Sepsis early warning device, equipment and storage medium |
| CN110993096A (en) * | 2019-11-27 | 2020-04-10 | 医惠科技有限公司 | Sepsis early warning device, equipment and storage medium |
| CN111081325A (en) * | 2019-12-27 | 2020-04-28 | 医渡云（北京）技术有限公司 | Medical data processing method and device |
| CN111110218A (en) * | 2019-12-31 | 2020-05-08 | 北京品驰医疗设备有限公司 | Sudden death from epilepsy prediction method and equipment |
| CN111326254A (en) * | 2018-12-17 | 2020-06-23 | 第四范式（北京）技术有限公司 | Method, apparatus, device and medium for index intervention |
| CN112205965A (en) * | 2020-08-28 | 2021-01-12 | 北京大学 | Health risk key event detection method and system based on time window cutting |
| CN112420201A (en) * | 2020-11-25 | 2021-02-26 | 哈尔滨工业大学 | Deep cascade framework for ICU mortality prediction and ICU mortality prediction method |
| CN112466469A (en) * | 2020-12-08 | 2021-03-09 | 杭州脉兴医疗科技有限公司 | Major crisis and death risk prediction method |
| CN112786166A (en) * | 2020-03-19 | 2021-05-11 | 中国医学科学院北京协和医院 | Data processing method, device, equipment and computer readable storage medium |
| CN112820368A (en) * | 2021-01-13 | 2021-05-18 | 中国人民解放军国防科技大学 | Method, system, device and storage medium for constructing critical patient data set |
| CN113270192A (en) * | 2021-04-19 | 2021-08-17 | 浙江大学 | Hemodialysis risk prediction system based on combination of dynamic and static data and depth self-encoder |
| CN116013495A (en) * | 2022-12-08 | 2023-04-25 | 广州视声健康科技有限公司 | Data monitoring method and device based on intelligent ward |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106021852A (en) * | 2016-05-06 | 2016-10-12 | 鼎泰生物科技（海南）有限公司 | Density clustering algorithm based blood glucose data abnormal value computing method and device |
| CN106361277A (en) * | 2016-08-26 | 2017-02-01 | 中山大学 | Sleep apnea syndrome assessment method based on electrocardiogram signals |
| CN106408001A (en) * | 2016-08-26 | 2017-02-15 | 西安电子科技大学 | Rapid area-of-interest detection method based on depth kernelized hashing |
| CN106446595A (en) * | 2016-12-16 | 2017-02-22 | 上海尚戴科技发展有限公司 | Gestational diabetes mellitus risk and degree prediction system based on machine learning |
| CN106446560A (en) * | 2016-09-28 | 2017-02-22 | 湖南老码信息科技有限责任公司 | Hyperlipidemia prediction method and prediction system based on incremental neural network model |
| CN106446765A (en) * | 2016-07-26 | 2017-02-22 | 重庆大学 | Health state evaluation system based on multidimensional physiological big data depth learning |
| CN106897545A (en) * | 2017-01-05 | 2017-06-27 | 浙江大学 | A kind of tumor prognosis forecasting system based on depth confidence network |
-
2017
- 2017-07-28 CN CN201710633561.9A patent/CN107491638A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106021852A (en) * | 2016-05-06 | 2016-10-12 | 鼎泰生物科技（海南）有限公司 | Density clustering algorithm based blood glucose data abnormal value computing method and device |
| CN106446765A (en) * | 2016-07-26 | 2017-02-22 | 重庆大学 | Health state evaluation system based on multidimensional physiological big data depth learning |
| CN106361277A (en) * | 2016-08-26 | 2017-02-01 | 中山大学 | Sleep apnea syndrome assessment method based on electrocardiogram signals |
| CN106408001A (en) * | 2016-08-26 | 2017-02-15 | 西安电子科技大学 | Rapid area-of-interest detection method based on depth kernelized hashing |
| CN106446560A (en) * | 2016-09-28 | 2017-02-22 | 湖南老码信息科技有限责任公司 | Hyperlipidemia prediction method and prediction system based on incremental neural network model |
| CN106446595A (en) * | 2016-12-16 | 2017-02-22 | 上海尚戴科技发展有限公司 | Gestational diabetes mellitus risk and degree prediction system based on machine learning |
| CN106897545A (en) * | 2017-01-05 | 2017-06-27 | 浙江大学 | A kind of tumor prognosis forecasting system based on depth confidence network |
Non-Patent Citations (4)
| Title |
|---|
| ANTONIO BOSNJAK等: "Predicting Mortality of ICU Patients Using Statistics of Physiological Variables and Support Vector Machines", 《2012 COMPUTING IN CARDIOLOGY》 * |
| GUAN WANG等: "Automatic Image-Based Plant Disease Severity Estimation Using Deep Learning", 《COMPUTER INTELLIGENCE AND NEUROSCIENCE》 * |
| MARZYEH GHASSEMI等: "Predicting intervention onset in the ICU with switching state space models", 《HTTPS://WWW.NCBI.NLM.NIH.GOV/PMC/ARTICLES/PMC5543372/》 * |
| MWLISSA ACZON等: "Dynamic Mortality Risk Predictions in Pediatric Critical Care Using Recurrent Neural Networks", 《ARXIV》 * |
Cited By (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108182974A (en) * | 2017-12-27 | 2018-06-19 | 深圳和而泰数据资源与云技术有限公司 | Illness appraisal procedure, terminal device and computer-readable medium |
| CN108198617A (en) * | 2017-12-27 | 2018-06-22 | 深圳和而泰数据资源与云技术有限公司 | Illness appraisal procedure, terminal device and computer-readable medium |
| CN108182974B (en) * | 2017-12-27 | 2021-01-05 | 深圳和而泰数据资源与云技术有限公司 | Disease evaluation method, terminal device and computer readable medium |
| CN108073497A (en) * | 2018-01-29 | 2018-05-25 | 上海洞识信息科技有限公司 | A kind of multi objective unusual fluctuation analysis method based on data center's data acquisition platform |
| CN108073497B (en) * | 2018-01-29 | 2021-02-12 | 上海洞识信息科技有限公司 | Multi-index transaction analysis method based on data center data acquisition platform |
| CN108877924A (en) * | 2018-03-30 | 2018-11-23 | 杭州依图医疗技术有限公司 | A kind of asthma method of determining probability and device |
| 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 |
| CN109359785A (en) * | 2018-11-30 | 2019-02-19 | 安徽四创电子股份有限公司 | Forced quarantine drug rehabilitation releasing personnel based on big data relapse hypothesis analysis system |
| CN111326254A (en) * | 2018-12-17 | 2020-06-23 | 第四范式（北京）技术有限公司 | Method, apparatus, device and medium for index intervention |
| CN111326254B (en) * | 2018-12-17 | 2024-06-07 | 第四范式（北京）技术有限公司 | Method, device, equipment and medium for index intervention |
| CN109684320A (en) * | 2018-12-25 | 2019-04-26 | 清华大学 | The method and apparatus of monitoring data on-line cleaning |
| CN109684320B (en) * | 2018-12-25 | 2020-09-15 | 清华大学 | Method and equipment for online cleaning of monitoring data |
| CN109887599A (en) * | 2019-02-25 | 2019-06-14 | 中国计量大学 | A kind of tcm prescription curative effect deduction method neural network based |
| CN110473636B (en) * | 2019-08-22 | 2020-09-22 | 山东众阳健康科技集团有限公司 | Intelligent medical advice recommendation method and system based on deep learning |
| CN110473636A (en) * | 2019-08-22 | 2019-11-19 | 山东众阳健康科技集团有限公司 | Intelligent doctor's advice recommended method and system based on deep learning |
| CN110911011B (en) * | 2019-11-27 | 2021-10-19 | 医惠科技有限公司 | Sepsis early warning device, equipment and storage medium |
| CN110911011A (en) * | 2019-11-27 | 2020-03-24 | 医惠科技有限公司 | Sepsis early warning device, equipment and storage medium |
| CN110993096B (en) * | 2019-11-27 | 2021-04-09 | 医惠科技有限公司 | Sepsis early warning device, equipment and storage medium |
| CN110993096A (en) * | 2019-11-27 | 2020-04-10 | 医惠科技有限公司 | Sepsis early warning device, equipment and storage medium |
| CN111081325A (en) * | 2019-12-27 | 2020-04-28 | 医渡云（北京）技术有限公司 | Medical data processing method and device |
| CN111081325B (en) * | 2019-12-27 | 2023-12-12 | 医渡云（北京）技术有限公司 | Medical data processing method and device |
| CN111110218A (en) * | 2019-12-31 | 2020-05-08 | 北京品驰医疗设备有限公司 | Sudden death from epilepsy prediction method and equipment |
| CN111110218B (en) * | 2019-12-31 | 2024-03-08 | 北京品驰医疗设备有限公司 | Sudden epileptic death prediction method and device |
| CN112786166A (en) * | 2020-03-19 | 2021-05-11 | 中国医学科学院北京协和医院 | Data processing method, device, equipment and computer readable storage medium |
| CN112205965A (en) * | 2020-08-28 | 2021-01-12 | 北京大学 | Health risk key event detection method and system based on time window cutting |
| CN112420201A (en) * | 2020-11-25 | 2021-02-26 | 哈尔滨工业大学 | Deep cascade framework for ICU mortality prediction and ICU mortality prediction method |
| CN112466469A (en) * | 2020-12-08 | 2021-03-09 | 杭州脉兴医疗科技有限公司 | Major crisis and death risk prediction method |
| CN112820368A (en) * | 2021-01-13 | 2021-05-18 | 中国人民解放军国防科技大学 | Method, system, device and storage medium for constructing critical patient data set |
| CN113270192A (en) * | 2021-04-19 | 2021-08-17 | 浙江大学 | Hemodialysis risk prediction system based on combination of dynamic and static data and depth self-encoder |
| CN116013495A (en) * | 2022-12-08 | 2023-04-25 | 广州视声健康科技有限公司 | Data monitoring method and device based on intelligent ward |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107491638A (en) | A kind of ICU user's prognosis method and terminal device based on deep learning model | |
| CN107595243A (en) | A kind of illness appraisal procedure and terminal device | |
| Cicceri et al. | A deep learning approach for pressure ulcer prevention using wearable computing | |
| Forkan et al. | ViSiBiD: A learning model for early discovery and real-time prediction of severe clinical events using vital signs as big data | |
| CN109119130A (en) | A kind of big data based on cloud computing is health management system arranged and method | |
| CN107863147A (en) | The method of medical diagnosis based on depth convolutional neural networks | |
| Morshed et al. | Deep osmosis: Holistic distributed deep learning in osmotic computing | |
| CN108428477A (en) | Construction method and cloud medical system based on the twinborn cloud surgery simulation platform of number | |
| Gandhi et al. | An automated review of body sensor networks research patterns and trends | |
| CN115910319A (en) | Otology inquiry assisting method and device, electronic equipment and storage medium | |
| CN109411053B (en) | Old people action rehabilitation training management data model construction method | |
| Nanwal et al. | Green IoT and Big Data: Succeeding towards Building Smart Cities | |
| CN109770904A (en) | Monitoring method, device, computer equipment and the storage medium of apnea | |
| CN106618592A (en) | Remote blood glucose monitoring processing method and mobile terminal | |
| Fahim et al. | Knowledge-based decision support systems for personalized u-lifecare big data services | |
| CN110427870B (en) | Eye picture recognition method, target recognition model training method and device | |
| Munagala et al. | Optimised ensemble learning-based IoT-enabled heart disease monitoring system: an optimal fuzzy ranking concept | |
| Murthy et al. | Deepseas: Smartphone-based early ailment sensing using coupled lstm autoencoders | |
| Choi et al. | A multi-view learning approach to enhance automatic 12-lead ECG diagnosis performance | |
| KR102451102B1 (en) | Companion animal skin lesion classification system and method based on artificial intelligence using clinical and image information | |
| CN114678126A (en) | Disease tracking and predicting system | |
| EP3979257A1 (en) | Enabling utilization of massive data in a service | |
| Arivalahan et al. | Health Monitoring in IoT with Context-Aware Deep Convolutional Spiking Neural Network and Woodpecker Mating Algorithm | |
| KR102501793B1 (en) | Method for constructing dataset | |
| Bozkurt | Wearable technology: The current state of the art |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171219 |