CN110458355A - Event prediction method, apparatus, equipment and storage medium - Google Patents
Event prediction method, apparatus, equipment and storage medium Download PDFInfo
- Publication number
- CN110458355A CN110458355A CN201910736035.4A CN201910736035A CN110458355A CN 110458355 A CN110458355 A CN 110458355A CN 201910736035 A CN201910736035 A CN 201910736035A CN 110458355 A CN110458355 A CN 110458355A
- Authority
- CN
- China
- Prior art keywords
- event
- data
- predetermined time
- prediction model
- occurred
- 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 86
- 238000003860 storage Methods 0.000 title claims abstract description 29
- 238000012549 training Methods 0.000 claims abstract description 75
- 230000007787 long-term memory Effects 0.000 claims abstract description 15
- 230000006870 function Effects 0.000 claims description 49
- 239000011159 matrix material Substances 0.000 claims description 19
- 238000012545 processing Methods 0.000 claims description 18
- 238000005457 optimization Methods 0.000 claims description 13
- 230000015654 memory Effects 0.000 claims description 12
- 230000009897 systematic effect Effects 0.000 claims description 6
- 201000010099 disease Diseases 0.000 description 27
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 27
- 238000005516 engineering process Methods 0.000 description 16
- 239000003814 drug Substances 0.000 description 15
- 229940079593 drug Drugs 0.000 description 14
- 238000004140 cleaning Methods 0.000 description 8
- 238000010276 construction Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 7
- 238000009826 distribution Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 6
- 238000013473 artificial intelligence Methods 0.000 description 5
- 230000000474 nursing effect Effects 0.000 description 5
- 230000006854 communication Effects 0.000 description 4
- 230000014759 maintenance of location Effects 0.000 description 4
- 208000024172 Cardiovascular disease Diseases 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 238000004590 computer program Methods 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 208000026106 cerebrovascular disease Diseases 0.000 description 2
- 230000002526 effect on cardiovascular system Effects 0.000 description 2
- 239000004744 fabric Substances 0.000 description 2
- 230000010365 information processing Effects 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 208000024827 Alzheimer disease Diseases 0.000 description 1
- 206010008190 Cerebrovascular accident Diseases 0.000 description 1
- 208000035473 Communicable disease Diseases 0.000 description 1
- 206010028980 Neoplasm Diseases 0.000 description 1
- 206010039966 Senile dementia Diseases 0.000 description 1
- 208000006011 Stroke Diseases 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 201000011510 cancer Diseases 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 208000015181 infectious disease Diseases 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 210000000867 larynx Anatomy 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000003014 reinforcing effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000009738 saturating Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Mathematical Physics (AREA)
- Economics (AREA)
- Tourism & Hospitality (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- General Business, Economics & Management (AREA)
- Educational Administration (AREA)
- Primary Health Care (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention discloses a kind of event prediction method, apparatus, equipment and storage medium.The event prediction method includes: to obtain the data of the event occurred within the first predetermined time；According to the data of the event occurred within first predetermined time, training set is determined；Based on the training set, shot and long term memory network is trained, obtains event prediction model；And according to the event prediction model, the event to be occurred in forecast date for the treatment of is predicted.
Description
Technical field
The present invention relates to intelligence endowment service technology fields, in particular to a kind of event prediction method, apparatus, set
Standby and storage medium.
Background technique
Currently, artificial intelligence is a branch of computer science, it attempts to understand the essence of intelligence, and produces one kind
The new intelligence machine that can be made a response in such a way that human intelligence is similar, the research in the field include robot, language identification,
Image recognition, natural language processing and expert system etc..For artificial intelligence since the birth, theory and technology is increasingly mature, application
Field also constantly expands.For example, artificial intelligence technology can be applied, intelligent event prediction is carried out.
Above- mentioned information are only used for reinforcing the understanding to background of the invention, therefore it disclosed in the background technology part
It may include the information not constituted to the prior art known to persons of ordinary skill in the art.
Summary of the invention
In view of this, the present invention provides a kind of event prediction method, apparatus, equipment and storage medium.
Other characteristics and advantages of the invention will be apparent from by the following detailed description, or partially by the present invention
Practice and acquistion.
According to an aspect of the present invention, a kind of event methods of being hospitalized are provided, comprising: acquisition occurs within the first predetermined time
Event data；According to the data of the event occurred within first predetermined time, training set is determined；Based on described
Training set is trained shot and long term memory network, obtains event prediction model；It is right and according to the event prediction model
The event to be occurred is predicted in date to be predicted.
According to an embodiment of the present invention, the method also includes: acquisition occur within second scheduled time it is described
The data of event；According to the data of the event occurred within second predetermined time, the event prediction model is determined
Verifying collection；Collected based on the verifying, place is iterated to the model parameter in the event prediction model by loss function
Reason；After each iteration updates, the value of loss function of the current event prediction model on the verifying collection is calculated；According to the damage
The value for losing function draws loss function curve；And when the loss function curve convergence, according to current event prediction model
Model Parameter Optimization described in event prediction model model parameter, to optimize the event prediction model.
According to an embodiment of the present invention, the event includes: endowment event；According within first predetermined time
The data of the event occurred determine that the training set of event prediction model includes: that basis is sent out within first predetermined time
The data of the raw endowment event determine the endowment event that every preset interval occurs within first predetermined time respectively
First number；And it regard described first time number as the training set.
According to an embodiment of the present invention, it is based on the training set, being trained to shot and long term memory network includes: root
Time-based matrix is constructed according to described first time number；And using the matrix as the training set, the length is remembered
Network is trained.
According to an embodiment of the present invention, according to the data of the event occurred within second predetermined time,
The verifying collection for determining the event prediction model includes: according to the endowment event occurred within second predetermined time
Data determine second number of the endowment event occurred in second predetermined time per the preset interval respectively；And
By described second time number as verifying collection.
According to an embodiment of the present invention, the endowment event includes: room cleaning event, disease emergency event, falls
The all or part in event and self-help shopping event.
According to an embodiment of the present invention, the event includes: event in hospital；According within first predetermined time
The data of the event occurred, determine event prediction model training set include: will be within first predetermined time per pre-
If the data for the event of being hospitalized that occurred at intervals are as the training set.
According to an embodiment of the present invention, it is based on the training set, being trained to shot and long term memory network includes: root
Time-based matrix is constructed according to the data of the event of being hospitalized；And using the matrix as the training set, to the length
Short memory network is trained.
According to an embodiment of the present invention, according to the data of the event occurred within second predetermined time,
The verifying collection for determining the event prediction model includes: to live what is occurred within second predetermined time per the preset interval
The data of institute's event collect as the verifying.
According to an embodiment of the present invention, the data for being hospitalized event include: that ward occupancy rate, each department are hospitalized people
Number, major disease distributive law, operation number of units and rare drug usage amount all or part.
According to an embodiment of the present invention, obtaining the data of the event occurred within the first predetermined time includes:
Its data for being stored in the event occurred in the first predetermined time is obtained from the block chain network constructed.
According to another aspect of the present invention, a kind of event prediction device is provided, comprising: data acquisition module, for obtaining
The data of the event occurred within the first predetermined time；Training set determining module, for according within first predetermined time
The data of the event occurred, determine the training set of event prediction model；Systematic training module, for being based on the training
Collection, is trained shot and long term memory network, obtains the event prediction model；And event prediction module, for according to institute
Event prediction model is stated, the event to be occurred in forecast date for the treatment of is predicted.
In accordance with a further aspect of the present invention, a kind of computer equipment is provided, comprising: memory, processor and be stored in institute
The executable instruction that can be run in memory and in the processor is stated, the processor executes real when the executable instruction
Now such as any one of the above method.
In accordance with a further aspect of the present invention, a kind of computer readable storage medium is provided, being stored thereon with computer can hold
Row instruction, realizes such as any one of the above method when the executable instruction is executed by processor.
Event prediction method according to the present invention, it is very good possessed by the information stored such as block chain technology to can use
Secret protection effect and its have open and clear, traceable and be not easy the features such as distorting, to a large amount of event case information
It is stored.And the event case information based on these storages, event prediction model is determined, so as to according to the event prediction mould
Type carries out intelligent predicting to the data for the event that may occur.
It should be understood that the above general description and the following detailed description are merely exemplary, this can not be limited
Invention.
Detailed description of the invention
Its example embodiment is described in detail by referring to accompanying drawing, above and other target of the invention, feature and advantage will
It becomes more fully apparent.
Fig. 1 is a kind of flow chart of event prediction method shown according to an illustrative embodiments.
Fig. 2 is the flow chart of another event prediction method shown according to an illustrative embodiments.
Fig. 3 is the flow chart according to another event prediction method shown in an illustrative embodiments.
Fig. 4 is the flow chart according to another event prediction method shown in an illustrative embodiments.
Fig. 5 is the flow chart according to another event prediction method shown in an illustrative embodiments.
Fig. 6 is the flow chart according to another event prediction method shown in an illustrative embodiments.
Fig. 7 is the flow chart according to another event prediction method shown in an illustrative embodiments.
Fig. 8 is the flow chart according to another event prediction method shown in an illustrative embodiments.
Fig. 9 is a kind of block diagram of event prediction device shown according to an illustrative embodiments.
Figure 10 is the structural schematic diagram according to a kind of electronic equipment shown in an illustrative embodiments.
Figure 11 is a kind of schematic diagram of computer readable storage medium shown according to an illustrative embodiments.
Specific embodiment
Example embodiment is described more fully with reference to the drawings.However, example embodiment can be with a variety of shapes
Formula is implemented, and is not understood as limited to example set forth herein；On the contrary, thesing embodiments are provided so that the present invention will more
Fully and completely, and by the design of example embodiment comprehensively it is communicated to those skilled in the art.Attached drawing is only the present invention
Schematic illustrations, be not necessarily drawn to scale.Identical appended drawing reference indicates same or similar part in figure, thus
Repetition thereof will be omitted.
In addition, described feature, structure or characteristic can be incorporated in one or more implementations in any suitable manner
In mode.In the following description, many details are provided to provide and fully understand to embodiments of the present invention.So
And it will be appreciated by persons skilled in the art that technical solution of the present invention can be practiced and omit one in the specific detail
Or more, or can be using other methods, constituent element, device, step etc..In other cases, it is not shown in detail or describes
Known features, method, apparatus, realization or operation are to avoid a presumptuous guest usurps the role of the host and each aspect of the present invention is made to thicken.
In addition, term " first ", " second " are used for descriptive purposes only and cannot be understood as indicating or suggesting relative importance
Or implicitly indicate the quantity of indicated technical characteristic.Define " first " as a result, the feature of " second " can be expressed or
Implicitly include one or more of the features.In the description of the present invention, the meaning of " plurality " is at least two, such as two
It is a, three etc., unless otherwise specifically defined.
Fig. 1 is a kind of flow chart of event prediction method shown according to an illustrative embodiments.
With reference to Fig. 1, event prediction method 10 includes:
In step s 102, the data of the event occurred within the first predetermined time are obtained.
The event for example can be with nest egg event, such as the living habit of old man, rule, disease, burst thing in community of supporting parents
The events such as part and demand.
The event can also be event of being hospitalized, and the data of event can be for example inpatient's hospital bed occupancy rate, again
Big Disease Distribution, operation number of units and rare drug usage amount etc..
In some embodiments, for example, can be obtained from the block chain network constructed its storage in the first pre- timing
The data of the event of interior generation.It, can be to the data of a large amount of above-mentioned events uploaded from its different node based on block chain technology
It is stored.Block chain technology is to verify to know together with storing data, using distributed node using block linked data structure
Algorithm guarantees the safety of data transmission and access to generate, using by automation foot with more new data, in the way of cryptography
The intelligent contract of this code composition programs the completely new distributed basis framework and calculation of one kind with operation data.In area
In the most of application scenarios of block chain technology, combined using the chain transaction data of block chain Hash pointer, the Hash calculation that encryption is learned
Digital signature mechanism is learned with encryption and realizes multi-level evidence approval in process of exchange, between Lai Shixian Different Individual counterparty
Trust problem.By the information stored based on block chain technology, there is extraordinary secret protection effect, and have open saturating
It is bright, traceable and be not easy the features such as distorting, there is very high compatible degree with the data information of intelligent event, therefore can be by area
Block chain technology is organically combined with intelligent event prediction method.
But invention is not limited thereto, can also store in equipment and obtain from other storage servers, Cloud Server etc.
State the data of event.
In step S104, according to the data of the event occurred within the first predetermined time, training set is determined.
First predetermined time can be for example 1 year, half a year, three months etc., and invention is not limited thereto, when concrete application
It can be configured according to actual needs.
In step s 106, it is based on training set, to shot and long term memory network (Long Short-Term Memory, LSTM)
It is trained, obtains event prediction model.
LSTM network is a kind of Recognition with Recurrent Neural Network, is suitable for being spaced and postponing relatively in processing and predicted time sequence
Long critical event.LSTM network is the neural network of current comparative maturity, and details are not described herein.
In step S108, according to event prediction model, event to be occurred in forecast date for the treatment of is predicted.
The event prediction method that embodiment provides according to the present invention can use the information institute stored such as block chain technology
The extraordinary secret protection effect having and its have open and clear, traceable and it is not easy the features such as distorting, to a large amount of
Event case information is stored.And the event case information based on these storages, event prediction model is determined, so as to basis
The event prediction model carries out intelligent predicting to the data for the event that may occur.
It will be clearly understood that the present disclosure describe how being formed and using particular example, but the principle of the present invention is not limited to
These exemplary any details.On the contrary, the introduction based on present disclosure, these principles can be applied to many other
Embodiment.
Fig. 2 is the flow chart of another event prediction method shown according to an illustrative embodiments.With it is shown in FIG. 1
Method 10 the difference is that, method 20 shown in Fig. 2 may further comprise:
In step S202, the data of the event occurred within second scheduled time are obtained.
Similarly, the second predetermined time may be 1 year, half a year, three months etc., invention is not limited thereto.
In step S204, according to the data of the event occurred within second scheduled time, event prediction model is determined
Verifying collection.
For example, by taking above-mentioned first predetermined time and the second predetermined time are 1 year as an example, available continuous 2 years
The data of event, and using the data of wherein First Year as above-mentioned training set, collect the data of second year as verifying.
In step S206, based on verifying collection, changed by loss function to the model parameter in event prediction model
Generation processing.
Loss function (Loss Function) is that chance event or its value in relation to stochastic variable are mapped as non-negative reality
Count the function to indicate " risk " or " loss " of the chance event.In the application, loss function usually as learning criterion with
Optimization problem is associated, i.e., by minimizing loss function solution and assessment models.Such as the quilt in statistics and machine learning
Parameter Estimation for model.
In step S208, after each iteration updates, loss function of the current event prediction model on verifying collection is calculated
Value.
In step S210, loss function curve is drawn according to the value of loss function.
In step S212, when loss function curve convergence, according to the Model Parameter Optimization of current event prediction model
The model parameter of event prediction model, to optimize event prediction model.
When loss function curve convergence, the optimal solution of the model parameter of event prediction model can be obtained, so that it is determined that going out
Optimal event prediction model.
The event prediction method that embodiment provides according to the present invention, can further verify event prediction method,
Obtain the optimal solution of model parameter.
Fig. 3 is the flow chart according to another event prediction method shown in an illustrative embodiments.Shown in Fig. 3
In method 30, event is illustrated by taking endowment event as an example.
As shown in figure 3, method 30 includes:
In step s 302, the data of the endowment event occurred within one first predetermined time are obtained.
As described above, it can be for example based on block chain technology, the number to a large amount of endowment events uploaded from its different node
According to relevant information stored.Wherein, endowment event for example may include: room cleaning, burst disease, tumble, self-service purchase
Object etc., the data for event of supporting parents for example may include: room cleaning number, the number of burst disease, the number of tumble and from
Help the number etc. of shopping.
First predetermined time for example can be 1 year, one month etc..In practical applications, it can set according to actual needs
Fixed, invention is not limited thereto.
It will be understood by those skilled in the art that can be a certain node (a certain for the data of a large amount of endowment events obtained
Endowment community) data of endowment event that occur within first predetermined time of interior old man, or old man in multiple nodes
The mean value etc. of the data of the endowment event occurred within first predetermined time, invention is not limited thereto.
In step s 304, according to the data of the endowment event occurred within the first predetermined time, first is determined respectively
First number of endowment event occurs for every preset interval in predetermined time.
According to the relevant information of the data of the endowment event occurred in the first predetermined time of acquisition, counts and determine
First number of endowment event occurs for every preset interval in first predetermined time.The preset interval can be set according to actual needs
It sets.
With the predetermined time for 1 year (12 months), preset interval is to count and determine 1 year for one day (24 hours)
In time, the number of each endowment event occurs daily.It is respectively room cleaning, burst disease, tumble, self-service purchase with the event of supporting parents
For object, count respectively the old men need daily room clear up number x2_i, burst disease (such as cardiovascular and cerebrovascular disease, in
The elderlys Chang Fa such as wind disease) number x3_i, tumble number x4_i and using self-help shopping machine self-help shopping number x5_
i.Wherein i is time variable namely x2_i, x3_i, x4_i and x5_i are time series.To count the predetermined time from 2018
December 31 1 day to 2018 January, preset interval are room clearing times X2=[x2_20180101, x2_ for 24 hours
20180102,x2_20180102,x2_20180103,…,x2_20181230,x2_20181231].Burst disease number X3,
Number of falls X4 and self-help shopping number X5 are as example, and details are not described herein.
In step S306, it regard first number as training set, shot and long term memory network is trained, to obtain event
Prediction model.
Through the above steps, after counting and determining above-mentioned first number that each endowment event occurs in preset interval,
It regard first number as training set, shot and long term memory network is trained, to obtain event prediction model.
Still with the above-mentioned predetermined time for 1 year (12 months), preset interval is one day (24 hours), and endowment event is respectively
For room cleaning, burst disease, tumble, self-help shopping, with date x1_i, i.e. X1=[20180101,20180102,
20180103,…,20181230,20181231].In conjunction with each first number determined through the above steps, when building is based on
Between matrix X=[X1, X2, X3, X4, X5], specifically,
Later, it using the matrix X of building as training set, inputs in LSTM network and is trained, to obtain event prediction
Model.
In model training, such as X [0] can be used as mode input, set 1 for time step, X [1:5] is made
For model output, model is trained, to obtain above-mentioned event prediction model.It is predicted using event prediction model
When, by predicted time input model, then the predicted value of each variable under the date can be calculated.
In some embodiments, each block chain node can upload the relevant information of the data of above-mentioned endowment event
It is stored into block chain network.It later, can be by the background server of the one high permission node in the block chain network
Come obtain all nodes in block chain network endowment event data relevant information, to carry out model training, to obtain
Above-mentioned event prediction model.In addition, block chain can also be uploaded to through model training event prediction model obtained
It is stored in the node of network.It stores the block chain node of the event prediction model and stores the data of above-mentioned endowment event
The block chain node of relevant information can be identical, be also possible to different.Each block chain node is equal in the block chain network
The event prediction model can be downloaded from block chain network, to carry out the prediction of the data of following endowment events.
In step S308, using the date to be predicted as the input of above-mentioned event prediction model, the date to be predicted is predicted
The frequency of interior endowment event.
Using the date to be predicted as the input of above-mentioned event prediction model, to predict event of respectively supporting parents in the date to be predicted
Frequency.
The event prediction method of embodiment according to the present invention, such as can use the information based on the storage of block chain technology
Possessed extraordinary secret protection effect and its have open and clear, traceable and it is not easy the features such as distorting, to a large amount of
The case information of endowment event stored.And the case information of the endowment event based on these storages, determine event prediction
Model, so as to according to the forecasting system to the endowment event of the endowment mechanisms old man such as community (such as living habit, rule,
Disease, emergency event, demand) frequency etc. carry out reasonable prediction, for the intelligence such as unmanned service centre of artificial intelligence
Management system proposes prediction scheme (such as emergency stocks of goods and materials, personnel) to old Man's Demands and carries out to the emergency event that may occur
Dynamic early-warning.
Fig. 4 is the flow chart according to another event prediction method shown in an illustrative embodiments.
The difference is that, event prediction method 40 shown in Fig. 4 is training outgoing event with method 30 shown in Fig. 3
After prediction model, also verifying optimization can be further carried out to event prediction model, obtain the optimal solution of model parameter.
With reference to Fig. 4, event prediction method 40 be can further include:
In step S402, the data of the endowment event occurred within second scheduled time are obtained.
Such as the number of the endowment event occurred within second scheduled time can be obtained from the block chain network constructed
According to.
Similarly, which may be 1 year, one month etc..
For example, by taking above-mentioned first predetermined time and the second predetermined time are 1 year as an example, available continuous 2 years
The data of endowment event, and using the first annual data therein as above-mentioned training set, it is tested the data of second year as following
Card collection etc..
Similarly, it will be understood by those skilled in the art that the data of a large amount of endowment events obtained can be a certain section
The data for the endowment event that point (a certain endowment community) interior old man occurs within second predetermined time, or multiple nodes
The mean value etc. of the data for the endowment event that interior old man occurs within second predetermined time, invention is not limited thereto.But it needs
Illustrate, when the data of the endowment event in the first predetermined time obtained in the above method 30 are old men in a certain node
Endowment event data when, the data of the endowment event in the second predetermined time obtained in the method also should be the node
The data of the endowment event of interior old man.And if the number of the endowment event in the first predetermined time obtained in the above method 30
According to the mean value of the data for the endowment event for being old man in multiple nodes, then supporting in the second predetermined time for obtaining in the method
The data of old event also should be the mean value of the data of the endowment event of old man in multiple node.
In step s 404, according to the data of the endowment event occurred within second scheduled time, second is determined respectively
Second number of endowment event occurs for every preset interval in predetermined time.
The preset interval is identical as the setting of preset interval in the above method 30.
Still with the second predetermined time be 1 year, preset interval be it is daily, endowment event include room cleaning, burst disease,
Fall, for self-help shopping, second number counted respectively include: the old men need room to clear up daily number x2 ' _ i,
Number x3 ' _ i of burst disease (such as cardiovascular and cerebrovascular disease, apoplexy the elderly Chang Fa disease), number x4 ' _ i of tumble and make
With number x5 ' _ i of self-help shopping machine self-help shopping.Wherein i is time variable namely x2 ' _ i, x3 ' _ i, x4 ' _ i and x5 ' _ i
It is time series.Each second time several X2 ', X3 ', the statistics of X4 ' and X5 ' are identical with the above method 30, and details are not described herein.
In step S406, by second number as verifying collection, by loss function to the model in event prediction model
Parameter is iterated processing.
Loss function is that chance event or its value in relation to stochastic variable are mapped as nonnegative real number to indicate that this is random
The function of " risk " or " loss " of event.In the application, loss function is associated usually as learning criterion with optimization problem,
I.e. by minimizing loss function solution and assessment models.Such as it is used for the parameter of model in statistics and machine learning and estimates
Meter.
In step S408, after each iteration updates, loss function of the current event prediction model on verifying collection is calculated
Value.
In step S410, loss function curve is drawn according to the value of loss function.
In step S412, when loss function curve convergence, according to the Model Parameter Optimization of current event prediction model
The model parameter of event prediction model, to optimize event prediction model.
When loss function curve convergence, the optimal solution of event prediction model parameter can be obtained, so that it is determined that going out most
Prediction model in excellent event prediction model.
The event prediction method that embodiment provides according to the present invention, further tests the prediction model trained
Card, obtains the optimal solution of model parameter.
Fig. 5 is the flow chart according to another event prediction method shown in an illustrative embodiments.
The difference is that, stored in the above-mentioned data by event with method 30 shown in Fig. 3 or method shown in Fig. 4 40
In the embodiment in block chain network, method 50 shown in fig. 5 before to model training to obtain event prediction model, into
One step provides the construction method of block chain network, comprising:
In step S502, using geriatric nursing home as node, the block chain network is constructed.
Above-mentioned geriatric nursing home can be for example each endowment operating agency in society, or may be one or more collection
The geriatric nursing home of group company subordinate.
In practical applications, which is defined with actual participation person's range of the block chain network, i.e., with reality
The member that registers in the block chain network system defines, and invention is not limited thereto.
In step S504, the relevant information of above-mentioned endowment event is stored with preset data store organisation.
The relevant information of above-mentioned endowment event specifically for example may include: the geriatric nursing home registered in block chain network
The unmanned service centre's operation of related endowment community's artificial intelligence of upload and maintenance experience is shared and management case information, it is self-service enter
Firmly (room cleaning is dashed forward for registration navigation intelligent robot emergency case case information, the common problem concerning life of the elderly and case information
Hair event etc.), the elderly fall down and protect case information, the elderly trip and real time communication interaction case information (video, sound
Frequently, children's dialogue etc.), the self-service purchase machine case information of the elderly's ordinary articles etc..Furthermore associated materials can also be will demonstrate that
Audio, video, image etc. are uploaded in block chain network.
For the relevant information of above-mentioned endowment event, can according to preset data store organisation, information storage means and
Agreement etc. stores, to guarantee the high efficiency of information storage and information processing.
Table 1 is showing according to data store organisation of the endowment event case in block chain network shown in an example
Example.
Table 1
Table 2 is the example according to data store organisation of the event prediction model in block chain network shown in an example.
Table 2
It should be noted that table 1 and table 2 are only of the invention as an example, not a limit.
Fig. 6 is the flow chart according to another event prediction method shown in an illustrative embodiments.Shown in Fig. 6
In method 60, event is illustrated by taking the event of being hospitalized as an example.
As shown in fig. 6, method 60 includes:
In step S602, the data for the event of being hospitalized that every preset interval within the first predetermined time occurs are obtained.
It is such as above-mentioned, such as it can be based on block chain technology, the data to a large amount of events in hospital uploaded from its different node
It is stored.Wherein, the data of event for example may include: ward occupancy rate, each department's number of hospitalized, major disease point in hospital
The all or part of cloth rate, operation number of units and rare drug usage amount.
First predetermined time for example can be 1 year, one month etc..In practical applications, it can set according to actual needs
Fixed, invention is not limited thereto.
It will be understood by those skilled in the art that can be a certain node (a certain for a large amount of data for being hospitalized events obtained
Hospital institution) event of being hospitalized of interior patient within first predetermined time data, or patient is at this in multiple nodes
The mean value etc. of the data of event of being hospitalized in first predetermined time, invention is not limited thereto.
In step s 604, using the data of event in hospital as training set, shot and long term memory network is trained, to obtain
Obtain event prediction model.
Using the data of the event of being hospitalized obtained through the above steps as training set, shot and long term memory network is instructed
Practice, to obtain event prediction model.
With the first predetermined time for 1 year (12 months), every preset interval is one day (24 hours), in hospital the data of event
Respectively ward occupancy rate x2_i, each department's number of hospitalized x3_i, major disease distributive law x4_i, operation number of units x5_i and rare
For drug usage amount x6_i.Wherein i is time variable namely x2_i, x3_i, x4_i, x5_i and x6_i are time series.
With the first predetermined time from December 31,1 day to 2018 January in 2018, preset interval is ward occupancy rate for 24 hours
X2=[x2_20180101, x2_20180102, x2_20180102, x2_20180103 ..., x2_20181230, x2_
20181231].Each department's number of hospitalized X3, major disease distributive law X4, operation number of units X5 and rare drug usage amount X6 are with this
For, details are not described herein.
The setting date be x1_i, i.e. X1=[20180101,20180102,20180103 ..., 20181230,
20181231], in conjunction with above-mentioned ward occupancy rate X2, each department's number of hospitalized X3, major disease distributive law X4, operation number of units X5
And rare drug usage amount X6 building is based on time series matrix X=[X1, X2, X3, X4, X5, X6], specifically,
Later, it using the matrix X of building as training set, inputs in LSTM network and is trained, to obtain event prediction mould
Type.
It can be mode input by X [t-1] in model training, X [t] is model output, and time step is for example set as
1, model is trained, to obtain above-mentioned event prediction model.Wherein t is chronomere, with above-mentioned preset interval phase
Close, if above-mentioned preset interval be one day, herein t also one day be unit.It is predicted using event prediction model
When, to predict each variate-value of day proxima luce (prox. luc) as mode input, the model output of forecasting system is prediction day inpatient
The predicted value of hospital bed occupancy rate, major disease distribution, operation number of units and rare drug usage amount etc..
In some embodiments, the relevant information of above-mentioned event of being hospitalized can be uploaded to block by each block chain node
It is stored in chain network.Later, it can be obtained by the background server of the one high permission node in the block chain network
The relevant information of the event of being hospitalized of all nodes in block chain network, to carry out model training, so that it is pre- to obtain above-mentioned event
Survey model.In addition, can also be uploaded in the node of block chain network through model training event prediction model obtained
It is stored.It stores the block chain node of the event prediction model and stores the block chain link of the relevant information of above-mentioned event of being hospitalized
Point can be identical, be also possible to different.Each block chain node can be from block chain network in the block chain network
The event prediction model is downloaded, to carry out the data prediction of following events of being hospitalized.
In step S606, using the data of the event of being hospitalized of the proxima luce (prox. luc) generation on date to be predicted as event prediction model
Input, predict event of being hospitalized to be occurred in the date to be predicted.
With the data of the event of being hospitalized of the proxima luce (prox. luc) on date to be predicted, (such as above-mentioned ward occupancy rate, each department are hospitalized people
Number, major disease distributive law, operation number of units and rare drug usage amount etc.), to predict the event of being hospitalized on the date to be predicted, such as
The related data of the event of being hospitalized on the date to be predicted.
The event prediction method of embodiment according to the present invention, such as can use the information based on the storage of block chain technology
Extremely have open and clear, traceable of possessed extraordinary secret protection effect and it is not easy the features such as distorting, to a large amount of
Be hospitalized event data stored.And the data based on these events of being hospitalized, event prediction model is determined, so as to basis
The forecasting system is predicted (such as distribution of inpatient's hospital bed occupancy rate, major disease, operation to the event of being hospitalized of medical institutions
Number of units and rare drug usage amount etc.) reasonable prediction is carried out, so that the data that medical institutions can go out according to model prediction, mention
It is preceding to carry out corresponding programming dispatching, such as the bed of hospital is disposed, dispatch associated specialist and medical worker prepare to hold a consultation or
Operation, while being supplemented etc. in time according to the service condition of drug.
Fig. 7 is the flow chart according to another event prediction method shown in an illustrative embodiments.
The difference is that, event prediction method 70 shown in Fig. 7 is training outgoing event with method 60 shown in fig. 6
After prediction model, verifying optimization can also be further carried out to event prediction model, to obtain the optimal solution of model parameter.
With reference to Fig. 7, event prediction method 70 be can further include:
In step S702, the data for the event of being hospitalized that every preset interval within second scheduled time occurs are obtained.
Such as living for every preset interval generation within second scheduled time can be obtained from the block chain network constructed
The data of institute's event.
Similarly, which may be 1 year, one month etc..
For example, by taking above-mentioned first predetermined time and the second predetermined time are 1 year as an example, available continuous 2 years
The data of event in hospital, and using the data of wherein First Year as above-mentioned training set, it is tested the data of second year as following
Card collection etc..
Similarly, it will be understood by those skilled in the art that a large amount of data for being hospitalized event obtained can be a certain section
The data of be hospitalized event of point (a certain hospital institution) the interior patient within second predetermined time, or sick in multiple nodes
The mean value etc. of the data of be hospitalized event of the people within second predetermined time, invention is not limited thereto.But it should be recognized that
When the data of the event of being hospitalized in the first predetermined time obtained in the above method 60 are the events of being hospitalized in a certain node
When data, the data of the event of being hospitalized in the second predetermined time for obtaining in this method 70 are should be the event of being hospitalized of the node
Data.
The preset interval is identical as the setting of preset interval in the above method 60.
Still with the second predetermined time be 1 year, preset interval be it is daily, the data of event are respectively ward occupancy rate in hospital
X2 ' _ i, each department's number of hospitalized x3 ' _ i, major disease distributive law x4 ' _ i, operation number of units x5 ' _ i and rare drug usage amount
For x6 ' _ i.Similarly, wherein being time sequence for time variable namely x2 ' _ i, x3 ' _ i, x4 ' _ i, x5 ' _ i and x6 ' _ i
Column.Each data X2 ', X3 ', X4 ', the statistics of X5 ' and X6 ' are identical as the above method 60, and details are not described herein.
In step S704, the data of the event of being hospitalized of every preset interval within second scheduled time are collected as verifying,
Processing is iterated to the model parameter in event prediction model by loss function.
Loss function is that chance event or its value in relation to stochastic variable are mapped as nonnegative real number to indicate that this is random
The function of " risk " or " loss " of event.In the application, loss function is associated usually as learning criterion with optimization problem,
I.e. by minimizing loss function solution and assessment models.Such as it is used for the parameter of model in statistics and machine learning and estimates
Meter.
In step S706, after each iteration updates, loss function of the current event prediction model on verifying collection is calculated
Value.
In step S708, loss function curve is drawn according to the value of loss function.
In step S710, when loss function curve convergence, according to the Model Parameter Optimization of current event prediction model
The model parameter of event prediction model, to optimize event prediction model.
When loss function curve convergence, the optimal solution of event prediction model parameter can be obtained, so that it is determined that going out most
Prediction model in excellent event prediction model.
The intelligent event prediction method that embodiment provides according to the present invention, can be further to pre- in event prediction model
It surveys model to be verified, obtains the optimal solution of model parameter.
Fig. 8 is the flow chart according to another intelligent event prediction method shown in an illustrative embodiments.
The difference is that, stored in the above-mentioned data by event with method 60 shown in fig. 6 or method shown in Fig. 7 70
In the embodiment in block chain network, method 80 shown in Fig. 8 before to model training to obtain event prediction model, into
One step provides the construction method of block chain network, comprising:
In step S802, using medical institutions as node, the block chain network is constructed.
Above-mentioned medical institutions can be for example each medical operating agency in society, or may be one or more collection
The geriatric nursing home of group company subordinate.
In practical applications, which is defined with actual participation person's range of the block chain network, i.e., with reality
The member that registers in the block chain network system defines, and invention is not limited thereto.
In step S804, the data of above-mentioned event of being hospitalized are stored with preset data store organisation.
The data of above-mentioned event of being hospitalized for example may include: the disease that the medical institutions registered in block chain network upload
The information sharing of people's Hospitalization situation and management case, hospital bed occupancy rate, (internal medicine, shell, nerve are outer for the distribution of inpatient department situation
Section, Gastroenterology dept., ear nose larynx, department of infectious disease, section etc.), inpatient's evaluation move in the major disease situation point of number of days, inpatient
Cloth (cancer, cardiovascular disease, senile dementia etc.), inpatient need how long the case where performing the operation distribution (is waited in line
Deng), the Usage of Medicine for Inpatients situation distribution (rare drug etc.) etc. hospital patients Hospitalization situation information and management information.In addition, also
Audio, video, the image etc. that can will demonstrate that associated materials are uploaded in block chain network.
It, can be according to preset data store organisation, information storage means and agreement for the data of above-mentioned event of being hospitalized
Etc. storing, to guarantee the high efficiency of information storage and information processing.
Table 3 be according to shown in an example outpatient's Hospitalization situation information and management information storage organization show
Example.
Table 3
Table 4 is showing for data store organisation of another event prediction model in block chain network shown according to an example
Example.
Table 4
It should be noted that table 3 and table 4 are only of the invention as an example, not a limit.
It will be appreciated by those skilled in the art that realizing that all or part of the steps of above embodiment is implemented as being held by CPU
Capable computer program.When the computer program is executed by CPU, execute above-mentioned defined by the above method provided by the invention
Function.The program can store in a kind of computer readable storage medium, which can be read-only memory,
Disk or CD etc..
Further, it should be noted that above-mentioned attached drawing is only according to the present invention included by the method for illustrative embodiments
Processing schematically illustrates, rather than limits purpose.It can be readily appreciated that above-mentioned processing shown in the drawings does not indicate or limits these
The time sequencing of processing.In addition, being also easy to understand, these processing, which can be, for example either synchronously or asynchronously to be executed in multiple modules
's.
Following is apparatus of the present invention embodiment, can be used for executing embodiment of the present invention method.For apparatus of the present invention reality
Undisclosed details in example is applied, embodiment of the present invention method is please referred to.
Fig. 9 is a kind of block diagram of event prediction device shown according to an illustrative embodiments.
With reference to Fig. 9, event prediction device 90 includes: data acquisition module 902, training set determining module 904, systematic training
Module 906 and event prediction module 908.
Wherein, data acquisition module 902 is used to obtain the data of the event occurred within the first predetermined time.
Training set determining module 904 is used for the data according to the event occurred within first predetermined time, really
Determine the training set of event prediction model.
Systematic training module 906 is used to be based on the training set, is trained to shot and long term memory network, obtains the thing
Part prediction model.
Event prediction module 908 is used to treat the thing to be occurred in forecast date according to the event prediction model
Part is predicted.
In some embodiments, device 90 further include: the second data acquisition module, verifying collection determining module, iterative processing
Module, functional value computing module, Drawing of Curve module and system optimization module.Wherein, the second data acquisition module is for obtaining
The data of the event occurred within second scheduled time.Verifying collection determining module is used for according in second predetermined time
The data of the event of interior generation determine the verifying collection of the event prediction model.Iterative processing module is used for based on described
Verifying collection, is iterated processing to the model parameter in the event prediction model by loss function.Functional value computing module
After updating for each iteration of iterative processing module, loss function of the current event prediction model on the verifying collection is calculated
Value.Drawing of Curve module is used to draw loss function curve according to the value of the loss function.System optimization module is for working as institute
When stating loss function curve convergence, according to the model of event prediction model described in the Model Parameter Optimization of current event prediction model
Parameter, to optimize the event prediction model.
In some embodiments, the event includes: endowment event；Training set determining module 904 includes: that first number is true
Order member and training set determination unit.Wherein, first several determination unit are used for according to occurring in first predetermined time
The data of the endowment event determine the endowment event that every preset interval occurs in first predetermined time respectively
First number.Training set determination unit is used to regard described first time number as the training set.
In some embodiments, the event includes: endowment event；Systematic training module 906 includes: matrix construction unit
And training set determination unit.Wherein, matrix construction unit is used to construct time-based matrix according to described first time number.Training
Collect determination unit to be used to be trained the length memory network using the matrix as the training set.
In some embodiments, the event includes: endowment event；Verifying collection determining module includes: that second number determines
Unit and verifying collection determination unit.Wherein, second several determination unit is used to be supported according to what is occurred in second predetermined time
The data of old event determine the of the endowment event occurred in second predetermined time per the preset interval respectively
Two numbers.Verifying collection determination unit is used for described second time number as verifying collection.
In some embodiments, the event includes: endowment event；The endowment event includes: room cleaning event, disease
All or part in sick emergency event, tumble event and self-help shopping event.
In some embodiments, the event includes: event in hospital；Training set determining module 904 includes: that training set determines
Unit.Wherein, training set determination unit is used for thing of being hospitalized described in preset interval every within first predetermined time generation
The data of part are as the training set.
In some embodiments, the event includes: event in hospital；Systematic training module 906 includes: matrix construction unit
And training set determination unit.Wherein, matrix construction unit is used to construct time-based square according to the data of the event of being hospitalized
Battle array.Training set determination unit is used to be trained the length memory network using the matrix as the training set.
In some embodiments, the event includes: event in hospital；Verifying collection determining module includes: that verifying collection is determining single
Member.The event of being hospitalized that wherein verifying collection determination unit is used to occur within second predetermined time per the preset interval
Data collect as the verifying.
In some embodiments, the event includes: event in hospital；The data for being hospitalized event include: that ward is moved in
Rate, each department's number of hospitalized, major disease distributive law, operation number of units and rare drug usage amount all or part.
In some embodiments, data acquisition module 902 is stored in for obtaining it from the block chain network constructed
The data of the event occurred in first predetermined time.
Event prediction device according to the present invention, it is very good possessed by the information stored such as block chain technology to can use
Secret protection effect and its have open and clear, traceable and be not easy the features such as distorting, to a large amount of event case information
It is stored.And the event case information based on these storages, event prediction model is determined, so as to according to the event prediction mould
Type carries out intelligent predicting to the data for the event that may occur.
It should be noted that above-mentioned block diagram shown in the drawings is functional entity, not necessarily must with physically or logically
Independent entity is corresponding.Can realize these functional entitys using software form, or in one or more hardware modules or
These functional entitys are realized in integrated circuit, or are realized in heterogeneous networks and/or processor device and/or microcontroller device
These functional entitys.
Figure 10 is the structural schematic diagram according to a kind of electronic equipment shown in an illustrative embodiments.It needs to illustrate
That the electronic equipment shown in Figure 10 is only an example, should not function to the embodiment of the present invention and use scope bring and appoint
What is limited.
As shown in Figure 10, electronic equipment 800 is showed in the form of general purpose computing device.The component packet of electronic equipment 800
It includes: at least one central processing unit (CPU) 801, it can be according to the program generation being stored in read-only memory (ROM) 802
Code executes various from the program code that at least one storage unit 808 is loaded into random access storage device (RAM) 803
Movement and processing appropriate.
Particularly, according to an embodiment of the invention, said program code can be executed by central processing unit 801, so that
Central processing unit 801 executes various exemplary implementations according to the present invention described in this specification above method embodiment part
The step of mode.For example, central processing unit 801 can execute the step as shown in FIG. 1 to FIG. 8.
In RAM 803, also it is stored with electronic equipment 800 and operates required various programs and data.CPU 801,ROM
802 and RAM 803 is connected with each other by bus 804.Input/output (I/O) interface 805 is also connected to bus 804.
I/O interface 805 is connected to lower component: the input unit 806 including keyboard, mouse etc.；It is penetrated including such as cathode
The output unit 807 of spool (CRT), liquid crystal display (LCD) etc. and loudspeaker etc.；Storage unit 808 including hard disk etc.；
And the communication unit 809 of the network interface card including LAN card, modem etc..Communication unit 809 via such as because
The network of spy's net executes communication process.Driver 810 is also connected to I/O interface 805 as needed.Detachable media 811, such as
Disk, CD, magneto-optic disk, semiconductor memory etc. are mounted on as needed on driver 810, in order to read from thereon
Computer program be mounted into storage unit 808 as needed.
Figure 11 is a kind of schematic diagram of computer readable storage medium shown according to an illustrative embodiments.
With reference to shown in Figure 11, the program product for being set as realizing the above method of embodiment according to the present invention is described
900, can using portable compact disc read only memory (CD-ROM) and including program code, and can in terminal device,
Such as it is run on PC.However, program product of the invention is without being limited thereto, in this document, readable storage medium storing program for executing can be with
To be any include or the tangible medium of storage program, the program can be commanded execution system, device or device use or
It is in connection.
Above-mentioned computer-readable medium carries one or more program, when said one or multiple programs are by one
When the equipment executes, so that the computer-readable medium realizes the function as shown in FIG. 1 to FIG. 8.
It is particularly shown and described exemplary embodiments of the present invention above.It should be appreciated that the present invention is unlimited
In detailed construction described herein, set-up mode or implementation method；On the contrary, it is intended to cover included in appended claims
Spirit and scope in various modifications and equivalence setting.
Claims (12)
1. a kind of event prediction method characterized by comprising
Obtain the data of the event occurred within the first predetermined time；
According to the data of the event occurred within first predetermined time, training set is determined；
Based on the training set, shot and long term memory network is trained, obtains event prediction model；And
According to the event prediction model, the event to be occurred in forecast date for the treatment of is predicted.
2. the method according to claim 1, wherein further include:
Obtain the data of the event occurred within second scheduled time；
According to the data of the event occurred within second predetermined time, the verifying of the event prediction model is determined
Collection；
Collected based on the verifying, processing is iterated to the model parameter in the event prediction model by loss function；
After each iteration updates, the value of loss function of the current event prediction model on the verifying collection is calculated；
Loss function curve is drawn according to the value of the loss function；And
When the loss function curve convergence, according to event prediction mould described in the Model Parameter Optimization of current event prediction model
The model parameter of type, to optimize the event prediction model.
3. method according to claim 1 or 2, which is characterized in that the event includes: endowment event；According to described
The data of the event occurred in first predetermined time, determine that the training set of event prediction model includes:
According to the data of the endowment event occurred within first predetermined time, determine respectively predetermined described first
First number of the endowment event that every preset interval occurs in the time；And
It regard described first time number as the training set.
4. according to the method described in claim 3, it is characterized in that, the training set is based on, to the progress of shot and long term memory network
Training includes:
Time-based matrix is constructed according to described first time number；And
Using the matrix as the training set, the length memory network is trained.
5. according to the method described in claim 3, it is characterized in that, according to the thing occurred within second predetermined time
The data of part determine that the verifying collection of the event prediction model includes:
According to the data of the endowment event occurred within second predetermined time, determine respectively predetermined described second
Second number of the endowment event occurred in the time per the preset interval；And
By described second time number as verifying collection.
6. method according to claim 1 or 2, which is characterized in that the event includes: event in hospital；According to described
The data of the event occurred in first predetermined time, determine event prediction model training set include: will be described first
The data for the event of being hospitalized that every preset interval occurs in predetermined time are as the training set.
7. according to the method described in claim 6, it is characterized in that, the training set is based on, to the progress of shot and long term memory network
Training includes:
Time-based matrix is constructed according to the data of the event of being hospitalized；And
Using the matrix as the training set, the length memory network is trained.
8. according to the method described in claim 6, it is characterized in that, according to the thing occurred within second predetermined time
The data of part, determine the event prediction model verifying collection include: by within second predetermined time per it is described it is default between
The data for event of being hospitalized described in the generation collect as the verifying.
9. the method according to claim 1, wherein obtaining the event occurred within the first predetermined time
Data include: that its number for being stored in the event occurred in the first predetermined time is obtained from the block chain network constructed
According to.
10. a kind of event prediction device characterized by comprising
Data acquisition module, for obtaining the data of the event occurred within the first predetermined time；
Training set determining module determines event for the data according to the event occurred within first predetermined time
The training set of prediction model；
Systematic training module is trained shot and long term memory network, obtains the event prediction for being based on the training set
Model；And
Event prediction module, for treating the event to be occurred in forecast date and carrying out according to the event prediction model
Prediction.
11. a kind of computer equipment, comprising: memory, processor and storage are in the memory and can be in the processor
The executable instruction of middle operation, which is characterized in that the processor realizes such as claim 1-9 when executing the executable instruction
Described in any item methods.
12. a kind of computer readable storage medium, is stored thereon with computer executable instructions, which is characterized in that described to hold
Row instruction realizes such as the described in any item methods of claim 1-9 when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910736035.4A CN110458355A (en) | 2019-08-09 | 2019-08-09 | Event prediction method, apparatus, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910736035.4A CN110458355A (en) | 2019-08-09 | 2019-08-09 | Event prediction method, apparatus, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110458355A true CN110458355A (en) | 2019-11-15 |
Family
ID=68485806
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910736035.4A Pending CN110458355A (en) | 2019-08-09 | 2019-08-09 | Event prediction method, apparatus, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110458355A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111402252A (en) * | 2020-04-02 | 2020-07-10 | 杨仙莲 | Accurate medical image analysis method and robot surgery system |
| CN113886435A (en) * | 2021-12-08 | 2022-01-04 | 国能大渡河大数据服务有限公司 | Information query method and system based on recurrent neural network |
| WO2022002108A1 (en) * | 2020-06-30 | 2022-01-06 | 5G Care Limited | Intelligent system and method for event tracking |
| CN115439206A (en) * | 2022-11-08 | 2022-12-06 | 税友信息技术有限公司 | Declaration data prediction method, device, equipment and medium |
| CN116050676A (en) * | 2023-03-31 | 2023-05-02 | 四川互慧软件有限公司 | Emergency plan recommendation system based on emergency group events |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180270269A1 (en) * | 2016-02-09 | 2018-09-20 | International Business Machines Corporation | Detecting and predicting cyber-attack phases in adjacent data processing environment regions |
| 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 |
| CN109242140A (en) * | 2018-07-24 | 2019-01-18 | 浙江工业大学 | A kind of traffic flow forecasting method based on LSTM_Attention network |
| CN109738939A (en) * | 2019-03-21 | 2019-05-10 | 蔡寅 | A kind of Precursory Observational Data method for detecting abnormality |
| CN109816008A (en) * | 2019-01-20 | 2019-05-28 | 北京工业大学 | A kind of astronomical big data light curve predicting abnormality method based on shot and long term memory network |
| CN109859469A (en) * | 2019-02-15 | 2019-06-07 | 重庆邮电大学 | A kind of vehicle flowrate prediction technique based on integrated LSTM neural network |
-
2019
- 2019-08-09 CN CN201910736035.4A patent/CN110458355A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180270269A1 (en) * | 2016-02-09 | 2018-09-20 | International Business Machines Corporation | Detecting and predicting cyber-attack phases in adjacent data processing environment regions |
| 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 |
| CN109242140A (en) * | 2018-07-24 | 2019-01-18 | 浙江工业大学 | A kind of traffic flow forecasting method based on LSTM_Attention network |
| CN109816008A (en) * | 2019-01-20 | 2019-05-28 | 北京工业大学 | A kind of astronomical big data light curve predicting abnormality method based on shot and long term memory network |
| CN109859469A (en) * | 2019-02-15 | 2019-06-07 | 重庆邮电大学 | A kind of vehicle flowrate prediction technique based on integrated LSTM neural network |
| CN109738939A (en) * | 2019-03-21 | 2019-05-10 | 蔡寅 | A kind of Precursory Observational Data method for detecting abnormality |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111402252A (en) * | 2020-04-02 | 2020-07-10 | 杨仙莲 | Accurate medical image analysis method and robot surgery system |
| WO2022002108A1 (en) * | 2020-06-30 | 2022-01-06 | 5G Care Limited | Intelligent system and method for event tracking |
| CN113886435A (en) * | 2021-12-08 | 2022-01-04 | 国能大渡河大数据服务有限公司 | Information query method and system based on recurrent neural network |
| CN115439206A (en) * | 2022-11-08 | 2022-12-06 | 税友信息技术有限公司 | Declaration data prediction method, device, equipment and medium |
| CN115439206B (en) * | 2022-11-08 | 2023-03-07 | 税友信息技术有限公司 | Declaration data prediction method, device, equipment and medium |
| CN116050676A (en) * | 2023-03-31 | 2023-05-02 | 四川互慧软件有限公司 | Emergency plan recommendation system based on emergency group events |
| CN116050676B (en) * | 2023-03-31 | 2023-06-16 | 四川互慧软件有限公司 | Emergency plan recommendation system based on emergency group events |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110458355A (en) | Event prediction method, apparatus, equipment and storage medium | |
| Betzel et al. | Specificity and robustness of long-distance connections in weighted, interareal connectomes | |
| Yousefi et al. | Chaotic genetic algorithm and Adaboost ensemble metamodeling approach for optimum resource planning in emergency departments | |
| Fong et al. | Composite Monte Carlo decision making under high uncertainty of novel coronavirus epidemic using hybridized deep learning and fuzzy rule induction | |
| Peak et al. | Comparing nonpharmaceutical interventions for containing emerging epidemics | |
| Garcia-Vicuña et al. | Hospital preparedness during epidemics using simulation: the case of COVID-19 | |
| Li et al. | MSGD: A novel matrix factorization approach for large-scale collaborative filtering recommender systems on GPUs | |
| Wen et al. | Fuzzy data envelopment analysis (DEA): Model and ranking method | |
| Yousefi et al. | Simulation-based optimization methods applied in hospital emergency departments: A systematic review | |
| Alipour-Vaezi et al. | Prioritizing and queueing the emergency departments’ patients using a novel data-driven decision-making methodology, a real case study | |
| Tasgetiren et al. | A variable block insertion heuristic for the blocking flowshop scheduling problem with total flowtime criterion | |
| Li et al. | Credibilistic programming | |
| Ning et al. | An uncertain multi-objective programming model for machine scheduling problem | |
| Xu et al. | Inexact two-stage stochastic robust optimization model for water resources management under uncertainty | |
| Garcia-Vicuña et al. | Planning ward and intensive care unit beds for COVID-19 patients using a discrete event simulation model | |
| CN112085378A (en) | Resource allocation method, device, computer equipment and storage medium | |
| Abdalkareem et al. | Discrete flower pollination algorithm for patient admission scheduling problem | |
| Zhang et al. | A new three-dimensional manufacturing service composition method under various structures using improved Flower Pollination Algorithm | |
| Zhang et al. | A unified framework for epidemic prediction based on poisson regression | |
| Kamiński et al. | On parallel policies for ranking and selection problems | |
| Abera et al. | On the decision support model for the patient admission scheduling problem with random arrivals and departures: A solution approach | |
| Abuhay et al. | Machine learning integrated patient flow simulation: why and how? | |
| Ding et al. | Two empirical uncertain models for project scheduling problem | |
| Park et al. | Designing a profit-maximizing critical peak pricing scheme considering the payback phenomenon | |
| Wu et al. | A meta-model-based multi-objective evolutionary approach to robust job shop scheduling |
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: 20191115 |