CN109739996A - A kind of construction method and device of industry knowledge mapping - Google Patents
A kind of construction method and device of industry knowledge mapping Download PDFInfo
- Publication number
- CN109739996A CN109739996A CN201811643488.4A CN201811643488A CN109739996A CN 109739996 A CN109739996 A CN 109739996A CN 201811643488 A CN201811643488 A CN 201811643488A CN 109739996 A CN109739996 A CN 109739996A
- Authority
- CN
- China
- Prior art keywords
- knowledge
- blocks
- deep learning
- data
- learning network
- 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.)
- Granted
Links
Landscapes
- Image Analysis (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention provides the construction methods and device of a kind of industrial knowledge mapping, wherein this method includes obtaining the feature vector of pending data；Establish simultaneously initialization algorithm model, wherein the algorithm model includes the first deep learning network and the second deep learning network；Based on the feature vector of the first deep learning network and the pending data, the first blocks of knowledge is generated；Based on the second deep learning network and first blocks of knowledge, the weighted value of the second blocks of knowledge and first blocks of knowledge and second blocks of knowledge is generated；According to the weighted value, the industrial knowledge mapping of the digraph comprising being directed toward second blocks of knowledge by first blocks of knowledge is generated.The embodiment of the present application forms industrial knowledge mapping, improves the systematicness and integrality of blocks of knowledge system by contacting between building blocks of knowledge and blocks of knowledge.
Description
Technical field
The present invention relates to technical field of data processing, a kind of construction method in particular to industrial knowledge mapping and
Device.
Background technique
A large amount of data can be generated in the operation or production process of equipment in industrial circle.For example, equipment is being transported
Row process, the clock signal or analog signal that the different sensors in equipment generate, the parameter of equipment itself is (such as equipment
Hardware address, the address of memory).If it is to be understood that the operation or the condition of production of equipment, will carry out at analysis above-mentioned data
Reason.
In the prior art, due to the complicated multiplicity of industrial data, and above-mentioned data are handled without corresponding model, is led
Cause can not form summary sex experience, and then can not establish contacting between other knowledge.
Summary of the invention
In view of this, the purpose of the present invention is to provide the construction methods and device of a kind of industrial knowledge mapping, with knowledge
The systematicness and integrality of unit system.
In a first aspect, the embodiment of the invention provides a kind of construction methods of industrial knowledge mapping, comprising:
Obtain the feature vector of pending data；
Establish simultaneously initialization algorithm model, wherein the algorithm model includes the first deep learning network and the second depth
Learning network；
Based on the feature vector of the first deep learning network and the pending data, the first blocks of knowledge is generated；
Based on the second deep learning network and first blocks of knowledge, the second blocks of knowledge and described first is generated
The weighted value of blocks of knowledge and second blocks of knowledge；
According to the weighted value, the digraph comprising being directed toward second blocks of knowledge by first blocks of knowledge is generated
Industrial knowledge mapping.
With reference to first aspect, the embodiment of the invention provides the first possible embodiments of first aspect, wherein institute
State the feature vector for obtaining pending data, comprising:
Establish data model, wherein the data model includes source data source, target side data source and the source
The mapping relations of data source and the target side data source；
Pending data is obtained from source by source data source；
Mapping relations based on the source data source and the target side data source obtain the pending data mapping
Target side data source, obtain the feature vector of the pending data.
The possible embodiment of with reference to first aspect the first, the embodiment of the invention provides second of first aspect
Possible embodiment, wherein the initialization algorithm model, comprising:
By the index configurations of the pending data in the first deep learning network and the second deep learning net
Network；
The training algorithm of the first deep learning network and the training algorithm of the second deep learning network are set.
With reference to first aspect, the embodiment of the invention provides the third possible embodiments of first aspect, wherein base
In the feature vector of the first deep learning network and the pending data, the first blocks of knowledge is generated, comprising:
The feature vector of the pending data is input to the first deep learning network, obtains model training knot
Fruit；
Model training result is input in knowledge base, makes the knowledge base according to the model training result and the mould
The mapping relations of type training result and blocks of knowledge generate the first blocks of knowledge.
With reference to first aspect, the embodiment of the invention provides the 4th kind of possible embodiments of first aspect, wherein institute
It states based on the second deep learning network and first blocks of knowledge, generates the second blocks of knowledge and the first knowledge list
First weighted value with second blocks of knowledge, comprising:
The feature vector of first blocks of knowledge and the pending data is input to the second deep learning net
The training layer of network obtains the second blocks of knowledge of the trained layer output；
First blocks of knowledge and second blocks of knowledge are input to the weight of the second deep learning network
It is worth generation layer, generates the weighted value of first blocks of knowledge and second blocks of knowledge.
Second aspect, the embodiment of the invention provides a kind of construction devices of industrial knowledge mapping, comprising:
Data acquisition module, for obtaining the feature vector of pending data；
Model building module, for establishing simultaneously initialization algorithm model, wherein the algorithm model includes the first depth
Practise network and the second deep learning network；
Knowledge mapping generation module, for the feature based on the first deep learning network and the pending data to
Amount generates the first blocks of knowledge, is based on the second deep learning network and first blocks of knowledge, generates the second knowledge list
The weighted value of first and described first blocks of knowledge and second blocks of knowledge is generated according to the weighted value comprising by described
First blocks of knowledge is directed toward the industrial knowledge mapping of the digraph of second blocks of knowledge.
In conjunction with second aspect, the embodiment of the invention provides the first possible embodiments of second aspect, wherein institute
Data acquisition module is stated, specifically for including:
Establish data model, wherein the data model includes source data source, target side data source and the source
The mapping relations of data source and the target side data source；
Pending data is obtained from source by source data source；
Mapping relations based on the source data source and the target side data source obtain the pending data mapping
Target side data source, obtain the feature vector of the pending data.
In conjunction with the first possible embodiment of second aspect, the embodiment of the invention provides second of second aspect
Possible embodiment, wherein the model building module, specifically for including:
By the index configurations of the pending data in the first deep learning network and the second deep learning net
Network；
The training algorithm of the first deep learning network and the training algorithm of the second deep learning network are set.
In conjunction with second aspect, the embodiment of the invention provides the third possible embodiments of second aspect, wherein institute
Knowledge mapping generation module is stated, specifically for including:
The feature vector of the pending data is input to the first deep learning network, obtains model training knot
Fruit；
Model training result is input in knowledge base, makes the knowledge base according to the model training result and the mould
The mapping relations of type training result and blocks of knowledge generate the first blocks of knowledge.
In conjunction with second aspect, the embodiment of the invention provides the 4th kind of possible embodiments of second aspect, wherein institute
State knowledge mapping generation module, be specifically also used to include:
The feature vector of first blocks of knowledge and the pending data is input to the second deep learning net
The training layer of network obtains the second blocks of knowledge of the trained layer output；
First blocks of knowledge and second blocks of knowledge are input to the weight of the second deep learning network
It is worth generation layer, generates the weighted value of first blocks of knowledge and second blocks of knowledge.
The construction method and device of industry knowledge mapping provided in an embodiment of the present invention, it is deep using pending data is inputted
Spend training in learning network, the industrial knowledge mapping of the digraph of the blocks of knowledge and blocks of knowledge of formation, and in the prior art
Can not establish blocks of knowledge compared with contacting between blocks of knowledge, improve the systematicness of blocks of knowledge system and complete
Property.
To enable the above objects, features and advantages of the present invention to be clearer and more comprehensible, preferred embodiment is cited below particularly, and cooperate
Appended attached drawing, is described in detail below.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be to needed in the embodiment attached
Figure is briefly described, it should be understood that the following drawings illustrates only certain embodiments of the present invention, therefore is not construed as pair
The restriction of range for those of ordinary skill in the art without creative efforts, can also be according to this
A little attached drawings obtain other relevant attached drawings.
Fig. 1 shows a kind of flow chart of the construction method of industry knowledge mapping provided by the embodiment of the present invention；
Fig. 2 shows the flow charts of the construction method of the industrial knowledge mapping of another kind provided by the embodiment of the present invention；
Fig. 3 shows the digraph that the first blocks of knowledge provided by the embodiment of the present invention is directed toward the second blocks of knowledge；
Fig. 4 shows a kind of structural schematic diagram of the construction device of industry knowledge mapping provided by the embodiment of the present invention.
Main element symbol description: data acquisition module 10；Model building module 11；Knowledge mapping generation module 12.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
Middle attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only
It is a part of the embodiment of the present invention, instead of all the embodiments.The present invention being usually described and illustrated herein in the accompanying drawings is real
The component for applying example can be arranged and be designed with a variety of different configurations.Therefore, of the invention to what is provided in the accompanying drawings below
The detailed description of embodiment is not intended to limit the range of claimed invention, but is merely representative of selected reality of the invention
Apply example.Based on the embodiment of the present invention, those skilled in the art institute obtained without making creative work
There are other embodiments, shall fall within the protection scope of the present invention.
Contacting between blocks of knowledge and blocks of knowledge can not be established in view of in the prior art, is based on this, the present invention
Embodiment provides the management method and device of a kind of database, is described below by embodiment.
The embodiment of the present application provides a kind of construction method of industrial knowledge mapping, including step S101-S105, specific to wrap
It includes:
Step S101 obtains the feature vector of pending data.
In the embodiment of the present application, pending data can be the operation data in industrial equipment, such as the sensing in equipment
Several in the data of device acquisition, the hardware address of equipment and device memory in the data such as data for storing.It is to be processed
Data can also be the parameter of part or product, by taking cutter as an example, can be size, material and sharpness of cutter etc.
Several in parameter.Pending data is multi-source heterogeneous data, can not directly as the data that algorithm model is directly handled, because
Pending data, is mapped to the form of feature vector by this.This implementation is provided for predicting cutting-tool's used life below
The construction method of industrial knowledge mapping be described.
As an alternative embodiment, the eigenvector method for obtaining pending data includes step S1010-
S1012, as shown in Fig. 2, specific as follows:
Step S1010, establishes data model, wherein data model includes source data source, target side data source and source
The mapping relations in end data source and target end data source.
Step S1011 obtains pending data from source by source data source.
Step S1012, the mapping relations based on source data source and target side data source obtain pending data mapping
Target side data source obtains the feature vector of pending data.
Specifically, the source data source of data model connects source, wherein when pending data refers to equipment, source can
Think memory, file or the database of storage device parameter, or acquisition component；Pending data nulling part or
When product, source can be storage product or memory, file or the database of Parameters of The Parts.From target side data source
The feature vector of generation can be obtained.The mapping relations of source data source and target side data source refer specifically to handle pending data
At the configuration of feature vector.Above-mentioned mapping relations are stored in the metadatabase of data model.When source data source is adopted from source
Collect pending data, metadata map component transfers above-mentioned mapping relations in metadatabase in data model, by pending data
It is mapped to feature vector.
Source can upload the file of the dimension information of the type cutter with user, edge length, the length of cutter hub such as cutter
Several parameter informations in the length of degree, the width of cutter hub and hilt.Source is also possible to the cutter stored in database
Material information, such as the material of cutter hub and the material of hilt.Source data source obtains above- mentioned information, the metadata in data model
Component is closed by the mapping relations of source data source and target side data source in calling metadatabase as a kind of possible mapping
System, the dimension information of all cutters corresponding characteristic value in above-mentioned mapping relations, as alternatively possible mapping relations,
The length of the length of cutter hub, the width of cutter hub and hilt a corresponding characteristic value, blade of cutter in above-mentioned mapping relations
Length corresponds to another characteristic value in above-mentioned mapping relations.By obtained characteristic value according to the sequential concatenation of setting, shape
At high-dimensional feature vector.
Step S102 establishes simultaneously initialization algorithm model, wherein algorithm model includes the first deep learning network and second
Deep learning network.
In the embodiment of the present application, feature vector is trained by the first deep learning network of algorithm model, shape
At knowledge (feature for extracting data), the second deep learning network that above-mentioned knowledge brings algorithm model into is trained, and is formed new
Knowledge (extracting new feature), and establish contacting between above-mentioned knowledge and above-mentioned new knowledge.
As an optional implementation manner, the method for initialization algorithm model includes (1) and (2), is specifically included:
(1) by the index configurations of pending data in the first deep learning network and the second deep learning network.
(2) training algorithm of the first deep learning network and the training algorithm of the second deep learning network are set.
Specifically, the index of pending data refer to pending data in the address of source, the first deep learning network and
The second available pending data of algorithm layer is trained, and forms knowledge.First deep learning network and the second algorithm layer obtain
The pending data taken can be different.In the page for configuring the first deep learning network, the index of pending data is configured, is passed through
It pulls algorithm and completes algorithm configuration.The configuration method of second deep learning network is identical, the algorithm of configuration and the first deep learning
The placement algorithm of network is different.
Step S103 generates the first blocks of knowledge based on the feature vector of the first deep learning network and pending data.
As an optional implementation manner, step S103 is specifically included as follows: the feature vector of pending data is defeated
Enter to the first deep learning network, obtains model training result.Model training result is input in knowledge base, knowledge base root is made
According to the mapping relations of model training result and model training result and blocks of knowledge, the first blocks of knowledge is generated.
Specifically, the first deep learning network is used to extract the feature of pending data, prediction result refers to number to be processed
According to assessment grade, the first blocks of knowledge refers to the corresponding evaluation of the assessment grade of pending data.Assessment cutting-tool's used life
Length in short-term, the parameter (several parameters such as size, sharpness of material and blade) for acquiring cutter is input to the first depth
Learning network is trained, and output indicates the symbol of the assessment grade of the length of cutting-tool's used life, such as opinion rating symbol
Including tetra- grades of A, B, C, D.It stores the corresponding evaluation of opinion rating in knowledge base, is grown very much as A corresponds to service life,
It is very longer that B corresponds to service life, and C corresponds to that service life is general, and it is short that A corresponds to service life.
Step S104 is based on the second deep learning network and the first blocks of knowledge, generates the second blocks of knowledge and first and knows
Know the weighted value of unit and the second blocks of knowledge.
As an alternative embodiment, the feature vector of the first blocks of knowledge and pending data is input to second
The training layer of deep learning network obtains the second blocks of knowledge of training layer output；By the first blocks of knowledge and the second knowledge list
Member is input to the weighted value generation layer of the second deep learning network, generates the weight of the first blocks of knowledge and the second blocks of knowledge
Value.
Specifically, the second blocks of knowledge refers to the new blocks of knowledge that the first blocks of knowledge combination pending data is formed.
The feature vector of the first blocks of knowledge generated in step S103 and pending data is input to the second deep learning network
Training layer is trained again, forms the second blocks of knowledge.By blocks of knowledge and the new blocks of knowledge (instruction of the second deep learning network
Practice the training result of layer) it is input to the weighted value generation layer of the second deep learning network, the second deep learning network exports weight
Value.Here weighted value indicates the tightness degree between two blocks of knowledge, and weighted value is bigger, the pass between two blocks of knowledge
It is closer.
After obtaining the evaluation of cutting-tool's used life, by the data and evaluation with cutter life length, input the
The training layer re -training of two deep learning networks, obtains more accurate evaluation, such as the influence factor of the length of service life.?
The influence factor of the length of the evaluation and service life of cutter life length is input to weight generation layer simultaneously, obtains cutter
The weighted value for contacting tightness degree of the influence factor of the length of the evaluation and service life of service life length, numerical value is bigger,
Tightness degree is higher.
As another optional embodiment, is input to by the first blocks of knowledge and with the feature vector of other data
The training layer of two deep learning networks obtains the second blocks of knowledge of training layer output.
After obtaining the evaluation of cutting-tool's used life, such as by the noise data of mechanical movement and above-mentioned evaluation, input the
The training layer re -training of two deep learning networks, the noise generated when obtaining using cutter.Cutter life length
Noise that evaluation and using cutter when generate while it being input to weight generation layer, obtain the evaluation of cutter life length and is made
The weighted value of the noise generated when with cutter.
Step S105 is generated according to weighted value comprising the digraph by the first blocks of knowledge the second blocks of knowledge of direction
Industrial knowledge mapping.
Specifically, the first blocks of knowledge and the second blocks of knowledge respectively represent the knowledge list after blocks of knowledge and training
Member.It is illustrated in figure 3 the digraph of the first blocks of knowledge and the second blocks of knowledge.Blocks of knowledge enabled node itself indicates, instructs
Blocks of knowledge after white silk is obtained by blocks of knowledge training, and the flow direction of knowledge is after blocks of knowledge flow direction training
Blocks of knowledge, usable direction indicate that the close relation of the blocks of knowledge after the big little finger of toe blocks of knowledge of weighted value and training can use company
The length of line indicates.According to the above method, the digraph of the blocks of knowledge after forming blocks of knowledge and training.Constantly basis is known
Unit knowledge after being trained is known, according to the oriented of the blocks of knowledge after weighted value building blocks of knowledge and training
Figure, can form knowledge mapping.User can choose the exhibition method of knowledge mapping, wherein exhibition method can be topological diagram, figure
If perhaps mind map exhibition method is topological diagram or mind map to mark, user can edit knowledge mapping by towing, compile
Knowledge mapping after volume can export.
The embodiment of the present application also provides a kind of construction device of industrial knowledge mapping, as shown in figure 4, including data acquisition mould
Block 10, model building module 11 and knowledge mapping generation module 12.
Data acquisition module 10, for obtaining the feature vector of pending data.
In the embodiment of the present application, pending data can be the operating parameter of equipment in industrial processes, with elevator
For, pending data can be the collected video of camera, the collected elevator interior of temperature sensor of elevator interior
Temperature and elevator the reflection elevator running state such as the speed of service data in several data.Pending data can also
The parameter for thinking part or product, by taking cutter as an example, pending data can be the cutting edge of a knife or a sword of the size of cutter, material and blade
Several parameters in the parameters such as sharp degree.The form of pending data can be constant parameter, or analog signal,
It can also be digital signal.Classification, form and the numerical value of pending data are typically different, usually multi-source heterogeneous data, nothing
Method directly carries out integration processing to them, therefore, first to form unified data structure.Feature vector is will be at pending data
The digital vectors obtained after reason.
As an alternative embodiment, data acquisition module by data model generate pending data feature to
Amount, specific as follows:
Establish data model, wherein data model include source data source, target side data source and source data source and
The mapping relations of target side data source；Pending data is obtained from source by source data source；Based on source data source and mesh
The mapping relations in mark end data source obtain the feature vector of pending data according to pending data.
Specifically, the source data source of data model connects source, wherein when pending data refers to equipment, source can
Think memory, file or the database of storage device parameter, or acquisition component；Pending data nulling part or
When product, source can be storage product or memory, file or the database of Parameters of The Parts.From target side data source
The feature vector of generation can be obtained.The mapping relations of source data source and target side data source are stored in the metadata of data model
In library.When source data source acquires pending data from source, metadata map component is transferred in metadatabase in data model
Pending data is mapped to feature vector by above-mentioned mapping relations.
Model building module 11, for establishing simultaneously initialization algorithm model, wherein algorithm model includes the first deep learning
Network and the second deep learning network.
In the embodiment of the present application, feature vector is trained by the first deep learning network of algorithm model, shape
At knowledge (feature for extracting data), the second deep learning network that above-mentioned knowledge brings algorithm model into is trained, and is formed new
Knowledge (extracting new feature), and establish contacting between above-mentioned knowledge and above-mentioned new knowledge.
It establishes after algorithm model, initialization algorithm model is also needed according to training, calculation is set in the process assignment of initialization
The parameter of method model.It is specific as follows:
By the index configurations of pending data in the first deep learning network and the second deep learning network.Setting first is deep
Spend the training algorithm of learning network and the training algorithm of the second deep learning network.
Specifically, the index of pending data refer to pending data in the address of source, the first deep learning network and
The second available pending data of algorithm layer is trained, and forms knowledge.First deep learning network and the second algorithm layer obtain
The pending data taken can be different.In the page for configuring the first deep learning network, the index of pending data is configured, is passed through
It pulls algorithm and completes algorithm configuration.The configuration method of second deep learning network is identical, the algorithm of configuration and the first deep learning
The placement algorithm of network is different.
Knowledge mapping generation module 12 generates digraph by executing (1), (2) and (3), specific as follows:
(1) feature vector based on the first deep learning network and pending data generates the first blocks of knowledge.
As an alternative embodiment, the feature vector of pending data is input to the first deep learning network,
Obtain model training result；Model training result is input in knowledge base, makes knowledge base according to model training result and model
The mapping relations of training result and blocks of knowledge generate the first blocks of knowledge.
Specifically, the first deep learning network is used to extract the feature of pending data, prediction result refers to number to be processed
According to assessment grade, blocks of knowledge refers to the evaluation of pending data.The length of assessment cutting-tool's used life in short-term, acquires cutter
Parameter (several parameters such as size, sharpness of material and blade) is input to the first deep learning network and is trained, defeated
The symbol of the assessment grade of the length of cutting-tool's used life is indicated out, and knowledge base obtains cutter by consulting symbol and evaluation
Service life length.
(2) it is based on the second deep learning network and the first blocks of knowledge, generates the second blocks of knowledge and the first blocks of knowledge
With the weighted value of the second blocks of knowledge.
As an alternative embodiment, the feature vector of the first blocks of knowledge and pending data is input to second
The training layer of deep learning network obtains the second blocks of knowledge of training layer output；By the first blocks of knowledge and the second knowledge list
Member is input to the weighted value generation layer of the second deep learning network, generates the weight of the first blocks of knowledge and the second blocks of knowledge
Value.
The feature vector of the blocks of knowledge generated in (1) and pending data is input to the instruction of the second deep learning network
Practice layer training (extracting feature) again, forms new blocks of knowledge.By blocks of knowledge and new blocks of knowledge (the second deep learning
The training result of the training layer of network) it is input to the weighted value generation layer of the second deep learning network, the second deep learning network
Export weighted value.Here weighted value indicates the tightness degree between two blocks of knowledge, and weighted value is bigger, two blocks of knowledge
Between relationship it is closer.
(3) according to weighted value, the industry for generating the digraph comprising being directed toward the second blocks of knowledge by the first blocks of knowledge is known
Know map.
Specifically, blocks of knowledge enabled node itself indicates, the blocks of knowledge after training is by blocks of knowledge, knowledge
Flow direction be from blocks of knowledge flow direction training after blocks of knowledge, usable direction indicate, the big little finger of toe blocks of knowledge of weighted value
With the close relation of the blocks of knowledge after training, can be indicated with the length of line.According to the above method, blocks of knowledge and instruction are formed
The digraph of blocks of knowledge after white silk.Constantly according to blocks of knowledge after being trained knowledge, according to weighted value structure
The digraph of blocks of knowledge after building blocks of knowledge and training, can form knowledge mapping.User can choose the exhibition of knowledge mapping
Show mode, wherein if exhibition method can be that perhaps mind map exhibition method is topological diagram or thinking for topological diagram, icon
Figure is led, user can edit knowledge mapping by towing, and edited knowledge mapping can export.
Based on above-mentioned analysis it is found that contacting phase with can not establish between blocks of knowledge and blocks of knowledge in the related technology
Than the construction method of industry knowledge mapping provided in an embodiment of the present invention, which uses, inputs pending data in deep learning network
Training, the industrial knowledge mapping of the digraph of the blocks of knowledge and blocks of knowledge of formation.
The computer program product that the construction method of industry knowledge mapping provided by the embodiment of the present invention carries out, including deposit
The computer readable storage medium of program code is stored up, the instruction that said program code includes can be used for executing previous methods implementation
Method described in example, specific implementation can be found in embodiment of the method, and details are not described herein.
The construction device of industry knowledge mapping provided by the embodiment of the present invention can in equipment specific hardware or
Software or firmware for being installed in equipment etc..The technology of device provided by the embodiment of the present invention, realization principle and generation is imitated
Fruit is identical with preceding method embodiment, and to briefly describe, Installation practice part does not refer to place, can refer to preceding method implementation
Corresponding contents in example.It is apparent to those skilled in the art that for convenience and simplicity of description, foregoing description
The specific work process of system, device and unit, the corresponding process during reference can be made to the above method embodiment, it is no longer superfluous herein
It states.
In embodiment provided by the present invention, it should be understood that disclosed device and method, it can be by others side
Formula is realized.The apparatus embodiments described above are merely exemplary, for example, the division of the unit, only one kind are patrolled
Function division is collected, there may be another division manner in actual implementation, in another example, multiple units or components can combine or can
To be integrated into another system, or some features can be ignored or not executed.Another point, shown or discussed is mutual
Coupling, direct-coupling or communication connection can be INDIRECT COUPLING or communication link by some communication interfaces, device or unit
It connects, can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
In addition, each functional unit in embodiment provided by the invention can integrate in one processing unit, it can also
To be that each unit physically exists alone, can also be integrated in one unit with two or more units.
It, can be with if the function is realized in the form of SFU software functional unit and when sold or used as an independent product
It is stored in a computer readable storage medium.Based on this understanding, technical solution of the present invention is substantially in other words
The part of the part that contributes to existing technology or the technical solution can be embodied in the form of software products, the meter
Calculation machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be a
People's computer, server or network equipment etc.) it performs all or part of the steps of the method described in the various embodiments of the present invention.
And storage medium above-mentioned includes: that USB flash disk, mobile hard disk, read-only memory (ROM, Read-OnlyMemory), arbitrary access are deposited
The various media that can store program code such as reservoir (RAM, Random Access Memory), magnetic or disk.
It should also be noted that similar label and letter indicate similar terms in following attached drawing, therefore, once a certain Xiang Yi
It is defined in a attached drawing, does not then need that it is further defined and explained in subsequent attached drawing, in addition, term " the
One ", " second ", " third " etc. are only used for distinguishing description, are not understood to indicate or imply relative importance.
Finally, it should be noted that embodiment described above, only a specific embodiment of the invention, to illustrate the present invention
Technical solution, rather than its limitations, scope of protection of the present invention is not limited thereto, although with reference to the foregoing embodiments to this hair
It is bright to be described in detail, those skilled in the art should understand that: anyone skilled in the art
In the technical scope disclosed by the present invention, it can still modify to technical solution documented by previous embodiment or can be light
It is readily conceivable that variation or equivalent replacement of some of the technical features；And these modifications, variation or replacement, do not make
The essence of corresponding technical solution is detached from the spirit and scope of technical solution of the embodiment of the present invention.Should all it cover in protection of the invention
Within the scope of.Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. a kind of construction method of industry knowledge mapping characterized by comprising
Obtain the feature vector of pending data；
Establish simultaneously initialization algorithm model, wherein the algorithm model includes the first deep learning network and the second deep learning
Network；
Based on the feature vector of the first deep learning network and the pending data, the first blocks of knowledge is generated；
Based on the second deep learning network and first blocks of knowledge, the second blocks of knowledge and first knowledge are generated
The weighted value of unit and second blocks of knowledge；
According to the weighted value, the work of the digraph comprising being directed toward second blocks of knowledge by first blocks of knowledge is generated
Industry knowledge mapping.
2. the method according to claim 1, wherein the feature vector for obtaining pending data, comprising:
Establish data model, wherein the data model includes source data source, target side data source and the source data
The mapping relations in source and the target side data source；
Pending data is obtained from source by source data source；
Mapping relations based on the source data source and the target side data source obtain the mesh of the pending data mapping
End data source is marked, the feature vector of the pending data is obtained.
3. according to the method described in claim 2, it is characterized in that, the initialization algorithm model, comprising:
By the index configurations of the pending data in the first deep learning network and the second deep learning network；
The training algorithm of the first deep learning network and the training algorithm of the second deep learning network are set.
4. the method according to claim 1, wherein based on the first deep learning network and described to be processed
The feature vector of data generates the first blocks of knowledge, comprising:
The feature vector of the pending data is input to the first deep learning network, obtains model training result；
Model training result is input in knowledge base, instructs the knowledge base according to the model training result and the model
Practice the mapping relations of result and blocks of knowledge, generates the first blocks of knowledge.
5. the method according to claim 1, wherein described based on the second deep learning network and described the
One blocks of knowledge generates the weighted value of the second blocks of knowledge and first blocks of knowledge and second blocks of knowledge, comprising:
The feature vector of first blocks of knowledge and the pending data is input to the second deep learning network
Training layer obtains the second blocks of knowledge of the trained layer output；
The weighted value that first blocks of knowledge and second blocks of knowledge are input to the second deep learning network is raw
Stratification generates the weighted value of first blocks of knowledge and second blocks of knowledge.
6. a kind of construction device of industry knowledge mapping characterized by comprising
Data acquisition module, for obtaining the feature vector of pending data；
Model building module, for establishing simultaneously initialization algorithm model, wherein the algorithm model includes the first deep learning net
Network and the second deep learning network；
Knowledge mapping generation module, for the feature vector based on the first deep learning network and the pending data,
The first blocks of knowledge is generated, the second deep learning network and first blocks of knowledge are based on, generates the second blocks of knowledge
It is generated comprising by described the with the weighted value of first blocks of knowledge and second blocks of knowledge according to the weighted value
One blocks of knowledge is directed toward the industrial knowledge mapping of the digraph of second blocks of knowledge.
7. device according to claim 6, which is characterized in that the data acquisition module, specifically for including:
Establish data model, wherein the data model includes source data source, target side data source and the source data
The mapping relations in source and the target side data source；
Pending data is obtained from source by source data source；
Mapping relations based on the source data source and the target side data source obtain the mesh of the pending data mapping
End data source is marked, the feature vector of the pending data is obtained.
8. device according to claim 7, which is characterized in that the model building module, specifically for including:
By the index configurations of the pending data in the first deep learning network and the second deep learning network；
The training algorithm of the first deep learning network and the training algorithm of the second deep learning network are set.
9. device according to claim 6, which is characterized in that the knowledge mapping generation module, specifically for including:
The feature vector of the pending data is input to the first deep learning network, obtains model training result；
Model training result is input in knowledge base, instructs the knowledge base according to the model training result and the model
Practice the mapping relations of result and blocks of knowledge, generates the first blocks of knowledge.
10. device according to claim 6, which is characterized in that the knowledge mapping generation module is specifically also used to wrap
It includes:
The feature vector of first blocks of knowledge and the pending data is input to the second deep learning network
Training layer obtains the second blocks of knowledge of the trained layer output；
The weighted value that first blocks of knowledge and second blocks of knowledge are input to the second deep learning network is raw
Stratification generates the weighted value of first blocks of knowledge and second blocks of knowledge.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811643488.4A CN109739996B (en) | 2018-12-29 | 2018-12-29 | Construction method and device of industrial knowledge map |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811643488.4A CN109739996B (en) | 2018-12-29 | 2018-12-29 | Construction method and device of industrial knowledge map |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109739996A true CN109739996A (en) | 2019-05-10 |
| CN109739996B CN109739996B (en) | 2020-12-25 |
Family
ID=66362601
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811643488.4A Active CN109739996B (en) | 2018-12-29 | 2018-12-29 | Construction method and device of industrial knowledge map |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109739996B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110957042A (en) * | 2020-01-17 | 2020-04-03 | 广州慧视医疗科技有限公司 | Prediction and simulation method of eye diseases under different conditions based on domain knowledge migration |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170103337A1 (en) * | 2015-10-08 | 2017-04-13 | International Business Machines Corporation | System and method to discover meaningful paths from linked open data |
| CN106933983A (en) * | 2017-02-20 | 2017-07-07 | 广东省中医院 | A kind of construction method of knowledge of TCM collection of illustrative plates |
| CN108550292A (en) * | 2018-04-16 | 2018-09-18 | 中山大学 | A kind of education resource multilayer tissue of on-line education system and representation method |
| CN108549937A (en) * | 2018-04-24 | 2018-09-18 | 厦门中控智慧信息技术有限公司 | A kind of knowledge migration method and device of detection network |
| CN108984745A (en) * | 2018-07-16 | 2018-12-11 | 福州大学 | A kind of neural network file classification method merging more knowledge mappings |
| CN109033129A (en) * | 2018-06-04 | 2018-12-18 | 桂林电子科技大学 | Multi-source Information Fusion knowledge mapping based on adaptive weighting indicates learning method |
-
2018
- 2018-12-29 CN CN201811643488.4A patent/CN109739996B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170103337A1 (en) * | 2015-10-08 | 2017-04-13 | International Business Machines Corporation | System and method to discover meaningful paths from linked open data |
| CN106933983A (en) * | 2017-02-20 | 2017-07-07 | 广东省中医院 | A kind of construction method of knowledge of TCM collection of illustrative plates |
| CN108550292A (en) * | 2018-04-16 | 2018-09-18 | 中山大学 | A kind of education resource multilayer tissue of on-line education system and representation method |
| CN108549937A (en) * | 2018-04-24 | 2018-09-18 | 厦门中控智慧信息技术有限公司 | A kind of knowledge migration method and device of detection network |
| CN109033129A (en) * | 2018-06-04 | 2018-12-18 | 桂林电子科技大学 | Multi-source Information Fusion knowledge mapping based on adaptive weighting indicates learning method |
| CN108984745A (en) * | 2018-07-16 | 2018-12-11 | 福州大学 | A kind of neural network file classification method merging more knowledge mappings |
Non-Patent Citations (2)
| Title |
|---|
| 杨晋吉 等: "一种知识图谱的排序学习个性化推荐算法", 《小型微型计算机系统》 * |
| 韩牧哲 等: "我国网络计量学研究的知识扩散可视化分析", 《图书情报研究》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110957042A (en) * | 2020-01-17 | 2020-04-03 | 广州慧视医疗科技有限公司 | Prediction and simulation method of eye diseases under different conditions based on domain knowledge migration |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109739996B (en) | 2020-12-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11783725B2 (en) | Snapshot management architecture for process control operator training system lifecycle | |
| US10345763B2 (en) | Cognitive plant clinic | |
| CN105717890B (en) | Industrial automation visualizes instrument board and creates example | |
| CN104914807B (en) | The incidence relation and alignment of determination process element and measured value in the process | |
| CN106796419A (en) | For the data pipe of Process Control System analysis | |
| CN104144204B (en) | A kind of method and system for the simulation for carrying out industrial automation system | |
| CN109685160B (en) | Online model automatic training and deploying method and system | |
| CN110198237A (en) | A kind of wireless configuration method for being intended to driving network | |
| CN110363449A (en) | A kind of Risk Identification Method, apparatus and system | |
| CN108564376A (en) | Risk control method, device, server and readable storage medium storing program for executing | |
| CN111523683B (en) | Method and system for predicting technological parameters in tobacco processing | |
| CN109639809A (en) | A kind of method and device of service data request link monitoring | |
| US20220253689A1 (en) | Predictive data capacity planning | |
| CN109816114A (en) | A kind of generation method of machine learning model, device | |
| CN108923983A (en) | Prediction technique, device and the readable storage medium storing program for executing of opportunistic network link | |
| CN114760669B (en) | Flow prediction-based route decision method and system | |
| CN113222170B (en) | Intelligent algorithm and model for AI collaborative service platform of Internet of things | |
| CN116127899B (en) | Chip design system, method, electronic device, and storage medium | |
| CN110300191A (en) | Service system and data processing method | |
| CN109032089A (en) | Data acquisition method and device for industrial equipment | |
| CN110414938A (en) | A kind of retrospect source system and method based on the building of platform of internet of things configurationization | |
| CN105680868B (en) | The method and apparatus of compressed stream data | |
| CN110299151A (en) | Detection method, detection model generation method and device | |
| CN109739996A (en) | A kind of construction method and device of industry knowledge mapping | |
| CN116455764A (en) | Digital twin network arranging method, digital twin network and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |