CN109739996B - Construction method and device of industrial knowledge map - Google Patents
Construction method and device of industrial knowledge map Download PDFInfo
- Publication number
- CN109739996B CN109739996B CN201811643488.4A CN201811643488A CN109739996B CN 109739996 B CN109739996 B CN 109739996B CN 201811643488 A CN201811643488 A CN 201811643488A CN 109739996 B CN109739996 B CN 109739996B
- Authority
- CN
- China
- Prior art keywords
- data
- knowledge
- processed
- deep learning
- 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.)
- Active
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a method and a device for constructing an industrial knowledge graph, wherein the method comprises the steps of obtaining a characteristic vector of data to be processed; establishing and initializing an algorithm model, wherein the algorithm model comprises a first deep learning network and a second deep learning network; generating a first knowledge unit based on the first deep learning network and the feature vector of the data to be processed; generating a second knowledge unit and weight values of the first knowledge unit and the second knowledge unit based on the second deep learning network and the first knowledge unit; and generating an industrial knowledge graph containing a directed graph pointed to by the first knowledge unit to the second knowledge unit according to the weight value. According to the embodiment of the application, the relation between the knowledge units is established to form the industrial knowledge map, so that the systematicness and the integrity of a knowledge unit system are improved.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a method and a device for constructing an industrial knowledge graph.
Background
In the industrial field, a large amount of data is generated during the operation or production of a device. For example, the device is running, timing signals or analog signals generated by different sensors on the device, and parameters of the device itself (e.g., hardware address of the device, address of the memory). If the operation or production condition of the equipment needs to be known, the data needs to be analyzed and processed.
In the prior art, due to the fact that industrial data are complex and various and no corresponding model is used for processing the data, summarized experience cannot be formed, and further relation with other knowledge cannot be established.
Disclosure of Invention
In view of the above, the present invention provides a method and an apparatus for constructing an industrial knowledge graph, which uses the systematicness and integrity of a knowledge unit system.
In a first aspect, an embodiment of the present invention provides a method for constructing an industrial knowledge graph, including:
acquiring a feature vector of data to be processed;
establishing and initializing an algorithm model, wherein the algorithm model comprises a first deep learning network and a second deep learning network;
generating a first knowledge unit based on the first deep learning network and the feature vector of the data to be processed;
generating a second knowledge unit and weight values of the first knowledge unit and the second knowledge unit based on the second deep learning network and the first knowledge unit;
and generating an industrial knowledge graph containing a directed graph pointed to by the first knowledge unit to the second knowledge unit according to the weight value.
With reference to the first aspect, an embodiment of the present invention provides a first possible implementation manner of the first aspect, where the obtaining a feature vector of to-be-processed data includes:
establishing a data model, wherein the data model comprises a source end data source, a target end data source and a mapping relation between the source end data source and the target end data source;
acquiring data to be processed from a source end through a source end data source;
and acquiring a target end data source mapped by the data to be processed based on the mapping relation between the source end data source and the target end data source, and acquiring the characteristic vector of the data to be processed.
With reference to the first possible implementation manner of the first aspect, an embodiment of the present invention provides a second possible implementation manner of the first aspect, where the initializing an algorithm model includes:
configuring the index of the data to be processed in the first deep learning network and the second deep learning network;
and setting a training algorithm of the first deep learning network and a training algorithm of the second deep learning network.
With reference to the first aspect, an embodiment of the present invention provides a third possible implementation manner of the first aspect, where generating a first knowledge unit based on the first deep learning network and the feature vector of the data to be processed includes:
inputting the feature vector of the data to be processed into the first deep learning network to obtain a model training result;
and inputting the model training result into a knowledge base, and enabling the knowledge base to generate a first knowledge unit according to the model training result and the mapping relation between the model training result and the knowledge unit.
With reference to the first aspect, an embodiment of the present invention provides a fourth possible implementation manner of the first aspect, where the generating weight values of a second knowledge unit and the first knowledge unit and the second knowledge unit based on the second deep learning network and the first knowledge unit includes:
inputting the first knowledge unit and the feature vector of the data to be processed into a training layer of the second deep learning network to obtain a second knowledge unit output by the training layer;
inputting the first knowledge unit and the second knowledge unit to a weight value generation layer of the second deep learning network, and generating weight values of the first knowledge unit and the second knowledge unit.
In a second aspect, an embodiment of the present invention provides an apparatus for constructing an industrial knowledge graph, including:
the data acquisition module is used for acquiring a feature vector of the data to be processed;
the model establishing module is used for establishing and initializing an algorithm model, wherein the algorithm model comprises a first deep learning network and a second deep learning network;
the knowledge graph generating module is used for generating a first knowledge unit based on the first deep learning network and the feature vector of the data to be processed, generating weight values of a second knowledge unit and the first knowledge unit and the second knowledge unit based on the second deep learning network and the first knowledge unit, and generating an industrial knowledge graph containing a directed graph directed to the second knowledge unit by the first knowledge unit according to the weight values.
With reference to the second aspect, an embodiment of the present invention provides a first possible implementation manner of the second aspect, where the data obtaining module is specifically configured to include:
establishing a data model, wherein the data model comprises a source end data source, a target end data source and a mapping relation between the source end data source and the target end data source;
acquiring data to be processed from a source end through a source end data source;
and acquiring a target end data source mapped by the data to be processed based on the mapping relation between the source end data source and the target end data source, and acquiring the characteristic vector of the data to be processed.
With reference to the first possible implementation manner of the second aspect, an embodiment of the present invention provides a second possible implementation manner of the second aspect, where the model establishing module is specifically configured to include:
configuring the index of the data to be processed in the first deep learning network and the second deep learning network;
and setting a training algorithm of the first deep learning network and a training algorithm of the second deep learning network.
With reference to the second aspect, an embodiment of the present invention provides a third possible implementation manner of the second aspect, where the knowledge graph generating module is specifically configured to include:
inputting the feature vector of the data to be processed into the first deep learning network to obtain a model training result;
and inputting the model training result into a knowledge base, and enabling the knowledge base to generate a first knowledge unit according to the model training result and the mapping relation between the model training result and the knowledge unit.
With reference to the second aspect, an embodiment of the present invention provides a fourth possible implementation manner of the second aspect, where the knowledge graph generating module is specifically further configured to:
inputting the first knowledge unit and the feature vector of the data to be processed into a training layer of the second deep learning network to obtain a second knowledge unit output by the training layer;
inputting the first knowledge unit and the second knowledge unit to a weight value generation layer of the second deep learning network, and generating weight values of the first knowledge unit and the second knowledge unit.
According to the method and the device for constructing the industrial knowledge graph, the data to be processed are input into the deep learning network for training, and the formed industrial knowledge graph of the directed graph of the knowledge units is formed.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.
FIG. 1 is a flow chart of a method for building an industrial knowledge graph according to an embodiment of the present invention;
FIG. 2 is a flow chart illustrating another method for building an industrial knowledge graph provided by an embodiment of the invention;
FIG. 3 illustrates a directed graph with a first knowledge unit pointing to a second knowledge unit provided by an embodiment of the invention;
FIG. 4 is a schematic structural diagram of an apparatus for building an industrial knowledge graph according to an embodiment of the present invention.
Description of the main element symbols: a data acquisition module 10; a model building module 11; a knowledge graph generation module 12.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without making any creative effort, shall fall within the protection scope of the present invention.
In view of the fact that the prior art cannot establish a relationship between knowledge units, embodiments of the present invention provide a method and an apparatus for managing a database, which are described below by way of embodiments.
The embodiment of the application provides a method for constructing an industrial knowledge graph, which comprises the following steps of S101-S105:
and step S101, acquiring a feature vector of the data to be processed.
In the embodiment of the present application, the data to be processed may be several pieces of data in the operation data of the industrial device, such as data collected by a sensor on the device, a hardware address of the device, and data stored in a memory of the device. The data to be processed may also be parameters of the part or product, and in the case of a tool, may be several of the parameters of the size, material, sharpness, etc. of the tool. The data to be processed is multi-source heterogeneous data and cannot be directly used as data directly processed by an algorithm model, so that the data to be processed is mapped into a characteristic vector form. The method for constructing the industrial knowledge map provided by the present embodiment will be described below by taking the prediction of the service life of the tool as an example.
As an optional implementation, the method for obtaining a feature vector of data to be processed includes steps S1010-S1012, as shown in fig. 2, specifically as follows:
step S1010, a data model is established, wherein the data model comprises a source end data source, a target end data source and a mapping relation between the source end data source and the target end data source.
Step S1011, obtaining the data to be processed from the source end through the source end data source.
Step S1012, obtaining a target end data source mapped by the data to be processed based on the mapping relationship between the source end data source and the target end data source, and obtaining a feature vector of the data to be processed.
Specifically, a source end data source of the data model is connected with a source end, wherein when the data to be processed refers to the device, the source end can be a memory, a file or a database for storing device parameters, and can also be an acquisition component; when the data to be processed refers to a part or a product, the source end may be a memory, a file, or a database storing parameters of the product or the part. The generated feature vector can be obtained from a target-end data source. The mapping relationship between the source data source and the target data source specifically refers to the configuration of processing the data to be processed into feature vectors. The mapping relationships are stored in a metadata base of the data model. When a source end data source collects data to be processed from a source end, a metadata mapping component in the data model calls the mapping relation in the metadata database, and the data to be processed is mapped into a feature vector.
The source end can upload a file of the size information of the type of the cutter by a user, such as a plurality of parameter information in the length of the cutting edge of the cutter, the length of the cutter body, the width of the cutter body and the length of the cutter handle. The source terminal may also be material information of the tool stored in the database, such as the material of the tool body and the material of the tool holder. The source end data source obtains the information, the metadata component in the data model is used as a possible mapping relation by calling the mapping relation between the source end data source and the target end data source in the metadata base, the size information of all the tools corresponds to one characteristic value in the mapping relation and is used as another possible mapping relation, the length of the tool body, the width of the tool body and the length of the tool handle correspond to one characteristic value in the mapping relation, and the length of the cutting edge of each tool corresponds to another characteristic value in the mapping relation. And splicing the obtained characteristic values according to a set sequence to form a high-dimensionality characteristic vector.
Step S102, an algorithm model is established and initialized, wherein the algorithm model comprises a first deep learning network and a second deep learning network.
In the embodiment of the application, the feature vector is trained through the first deep learning network of the algorithm model to form knowledge (extracting the features of data), the knowledge is brought into the second deep learning network of the algorithm model to be trained to form new knowledge (extracting the new features), and the relation between the knowledge and the new knowledge is established.
As an optional implementation manner, the method for initializing the algorithm model includes (1) and (2), and specifically includes:
(1) and configuring indexes of the data to be processed in the first deep learning network and the second deep learning network.
(2) And setting a training algorithm of the first deep learning network and a training algorithm of the second deep learning network.
Specifically, the index of the data to be processed refers to the address of the data to be processed at the source end, and the first deep learning network and the second algorithm layer can acquire the data to be processed for training to form knowledge. The data to be processed obtained by the first deep learning network and the second algorithm layer can be different. And configuring indexes of data to be processed on a page configured with the first deep learning network, and finishing algorithm configuration through a dragging algorithm. The configuration method of the second deep learning network is the same, and the configured algorithm is different from that of the first deep learning network.
Step S103, generating a first knowledge unit based on the first deep learning network and the feature vector of the data to be processed.
As an optional implementation manner, the step S103 specifically includes inputting the feature vector of the data to be processed into the first deep learning network, and obtaining the model training result. And inputting the model training result into a knowledge base, and enabling the knowledge base to generate a first knowledge unit according to the model training result and the mapping relation between the model training result and the knowledge unit.
Specifically, the first deep learning network is used for extracting features of the data to be processed, the prediction result refers to an evaluation level of the data to be processed, and the first knowledge unit refers to an evaluation corresponding to the evaluation level of the data to be processed. When the service life of the cutter is evaluated, parameters (a plurality of parameters such as size, material and sharpness of a blade edge) of the cutter are collected and input into the first deep learning network for training, and symbols of evaluation levels representing the service life of the cutter are output, wherein the evaluation level symbols comprise A, B, C, D four levels. The knowledge base stores the evaluation corresponding to the evaluation level, for example, A corresponds to very long service life, B corresponds to very long service life, C corresponds to general service life, and A corresponds to short service life.
And step S104, generating a second knowledge unit and weight values of the first knowledge unit and the second knowledge unit based on the second deep learning network and the first knowledge unit.
As an optional implementation manner, the first knowledge unit and the feature vector of the data to be processed are input to a training layer of a second deep learning network, and a second knowledge unit output by the training layer is obtained; and inputting the first knowledge unit and the second knowledge unit into a weight value generation layer of the second deep learning network to generate weight values of the first knowledge unit and the second knowledge unit.
In particular, the second knowledge unit refers to a new knowledge unit formed by the first knowledge unit in combination with the data to be processed. And inputting the first knowledge unit generated in the step S103 and the feature vector of the data to be processed into a training layer of a second deep learning network for training again to form a second knowledge unit. The knowledge unit and the new knowledge unit (training result of the training layer of the second deep learning network) are input to the weight value generation layer of the second deep learning network, and the second deep learning network outputs the weight value. The weight value here indicates the degree of closeness between two knowledge units, and the greater the weight value, the closer the relationship between two knowledge units.
And after the evaluation of the service life of the cutter is obtained, inputting the data and the evaluation of the service life of the cutter into a training layer of a second deep learning network for retraining, and obtaining more accurate evaluation, such as influence factors of the service life. And simultaneously inputting the evaluation of the service life of the cutter and the influence factor of the service life into the weight generation layer to obtain a weight value of the closeness degree of the connection between the evaluation of the service life of the cutter and the influence factor of the service life of the cutter, wherein the larger the numerical value, the higher the closeness degree is.
As another alternative, the first knowledge unit and the feature vector of other data are input to the training layer of the second deep learning network, and the second knowledge unit output by the training layer is obtained.
After the evaluation of the service life of the tool is obtained, noise data of machine operation and the evaluation are input into a training layer of a second deep learning network for retraining, and noise generated when the tool is used is obtained. And simultaneously inputting the evaluation of the service life of the cutter and the noise generated when the cutter is used into the weight generation layer to obtain the evaluation of the service life of the cutter and the weight value of the noise generated when the cutter is used.
And step S105, generating an industrial knowledge graph containing a directed graph pointing to the second knowledge unit from the first knowledge unit according to the weight value.
Specifically, the first knowledge unit and the second knowledge unit represent a knowledge unit and a trained knowledge unit, respectively. Fig. 3 shows a directed graph of a first knowledge unit and a second knowledge unit. The knowledge units can be represented by nodes, the trained knowledge units are obtained by training the knowledge units, the flow direction of knowledge flows from the knowledge units to the trained knowledge units and can be represented by directions, the weight values refer to the close relationship between the knowledge units and the trained knowledge units and can be represented by the length of connecting lines. According to the above method, a directed graph of knowledge units and trained knowledge units is formed. And training the knowledge units continuously to obtain trained knowledge, and constructing directed graphs of the knowledge units and the trained knowledge units according to the weight values to form knowledge graphs. The user can select the display mode of the knowledge graph, wherein the display mode can be a topological graph, an icon or a thought-derivative graph, and if the display mode is the topological graph or the thought-derivative graph, the user can edit the knowledge graph by dragging, and the edited knowledge graph can be exported.
The embodiment of the present application further provides a device for constructing an industrial knowledge graph, as shown in fig. 4, including a data obtaining module 10, a model building module 11, and a knowledge graph generating module 12.
And the data acquisition module 10 is used for acquiring the feature vector of the data to be processed.
In the embodiment of the application, the data to be processed may be operation parameters of equipment in an industrial production process, and taking an elevator as an example, the data to be processed may be a plurality of data in data reflecting an operation condition of the elevator, such as a video acquired by a camera inside the elevator, a temperature acquired by a temperature sensor inside the elevator, an operation speed of the elevator, and the like. The data to be processed may also be parameters of a part or a product, and in the case of a tool, the data to be processed may be several parameters of the size, the material, the sharpness of the edge, and the like of the tool. The form of the data to be processed may be a constant parameter, may be an analog signal, or may be a digital signal. The types, forms and values of the data to be processed are usually different, and usually are multi-source heterogeneous data, which cannot be directly integrated, so that a uniform data structure is formed first. The feature vector is a digital vector obtained by processing data to be processed.
As an optional implementation manner, the data obtaining module generates a feature vector of the data to be processed through the data model, which is specifically as follows:
establishing a data model, wherein the data model comprises a source end data source, a target end data source and a mapping relation between the source end data source and the target end data source; acquiring data to be processed from a source end through a source end data source; and acquiring a characteristic vector of the data to be processed according to the data to be processed based on the mapping relation between the source end data source and the target end data source.
Specifically, a source end data source of the data model is connected with a source end, wherein when the data to be processed refers to the device, the source end can be a memory, a file or a database for storing device parameters, and can also be an acquisition component; when the data to be processed refers to a part or a product, the source end may be a memory, a file, or a database storing parameters of the product or the part. The generated feature vector can be obtained from a target-end data source. The mapping relation between the source data source and the target data source is stored in a metadata base of the data model. When a source end data source collects data to be processed from a source end, a metadata mapping component in the data model calls the mapping relation in the metadata database, and the data to be processed is mapped into a feature vector.
And the model establishing module 11 is used for establishing and initializing an algorithm model, wherein the algorithm model comprises a first deep learning network and a second deep learning network.
In the embodiment of the application, the feature vector is trained through the first deep learning network of the algorithm model to form knowledge (extracting the features of data), the knowledge is brought into the second deep learning network of the algorithm model to be trained to form new knowledge (extracting the new features), and the relation between the knowledge and the new knowledge is established.
After the algorithm model is established, the algorithm model is initialized according to training requirements, and the initialization process refers to the configuration of parameters of the algorithm model. The method comprises the following specific steps:
and configuring indexes of the data to be processed in the first deep learning network and the second deep learning network. And setting a training algorithm of the first deep learning network and a training algorithm of the second deep learning network.
Specifically, the index of the data to be processed refers to the address of the data to be processed at the source end, and the first deep learning network and the second algorithm layer can acquire the data to be processed for training to form knowledge. The data to be processed obtained by the first deep learning network and the second algorithm layer can be different. And configuring indexes of data to be processed on a page configured with the first deep learning network, and finishing algorithm configuration through a dragging algorithm. The configuration method of the second deep learning network is the same, and the configured algorithm is different from that of the first deep learning network.
The knowledge graph generation module 12 generates a directed graph by performing (1), (2), and (3), specifically as follows:
(1) and generating a first knowledge unit based on the first deep learning network and the feature vector of the data to be processed.
As an optional implementation manner, inputting a feature vector of data to be processed into a first deep learning network to obtain a model training result; and inputting the model training result into a knowledge base, and enabling the knowledge base to generate a first knowledge unit according to the model training result and the mapping relation between the model training result and the knowledge unit.
Specifically, the first deep learning network is used for extracting features of the data to be processed, the prediction result refers to the evaluation level of the data to be processed, and the knowledge unit refers to the evaluation of the data to be processed. When the service life of the cutter is evaluated, parameters (a plurality of parameters such as size, material and sharpness of a blade) of the cutter are collected and input into the first deep learning network for training, a symbol of an evaluation level representing the service life of the cutter is output, and the knowledge base refers to the symbol and the evaluation to obtain the service life of the cutter.
(2) And generating the second knowledge unit and the weight values of the first knowledge unit and the second knowledge unit based on the second deep learning network and the first knowledge unit.
As an optional implementation manner, the first knowledge unit and the feature vector of the data to be processed are input to a training layer of a second deep learning network, and a second knowledge unit output by the training layer is obtained; and inputting the first knowledge unit and the second knowledge unit into a weight value generation layer of the second deep learning network to generate weight values of the first knowledge unit and the second knowledge unit.
And (3) inputting the knowledge unit generated in the step (1) and the feature vector of the data to be processed into a training layer of a second deep learning network for training again (extracting features), and forming a new knowledge unit. The knowledge unit and the new knowledge unit (training result of the training layer of the second deep learning network) are input to the weight value generation layer of the second deep learning network, and the second deep learning network outputs the weight value. The weight value here indicates the degree of closeness between two knowledge units, and the greater the weight value, the closer the relationship between two knowledge units.
(3) And generating an industrial knowledge graph containing a directed graph pointed to by the first knowledge unit to the second knowledge unit according to the weight values.
Specifically, the knowledge unit itself can be represented by a node, the trained knowledge unit passes through the knowledge unit, the flow direction of knowledge flows from the knowledge unit to the trained knowledge unit, the direction can be represented, the weight value refers to the close relationship between the knowledge unit and the trained knowledge unit, and the length of a connecting line can be represented. According to the above method, a directed graph of knowledge units and trained knowledge units is formed. And training the knowledge units continuously to obtain trained knowledge, and constructing directed graphs of the knowledge units and the trained knowledge units according to the weight values to form knowledge graphs. The user can select the display mode of the knowledge graph, wherein the display mode can be a topological graph, an icon or a thought-derivative graph, and if the display mode is the topological graph or the thought-derivative graph, the user can edit the knowledge graph by dragging, and the edited knowledge graph can be exported.
Based on the analysis, compared with the related technology that the relation between the knowledge unit and the knowledge unit cannot be established, the method for constructing the industrial knowledge graph provided by the embodiment of the invention adopts the industrial knowledge graph of the directed graph of the knowledge unit and the knowledge unit, which is formed by inputting the data to be processed into the deep learning network for training.
The computer program product of the method for constructing an industrial knowledge graph provided in the embodiment of the present invention includes a computer-readable storage medium storing a program code, where instructions included in the program code may be used to execute the method described in the foregoing method embodiment, and specific implementation may refer to the method embodiment, and will not be described herein again.
The device for constructing the industrial knowledge graph provided by the embodiment of the invention can be specific hardware on equipment, or software or firmware installed on the equipment. The device provided by the embodiment of the present invention has the same implementation principle and technical effect as the method embodiments, and for the sake of brief description, reference may be made to the corresponding contents in the method embodiments without reference to the device embodiments. It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the foregoing systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and there may be other divisions when actually implemented, and for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of devices or units through some communication interfaces, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments provided by the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus once an item is defined in one figure, it need not be further defined and explained in subsequent figures, and moreover, the terms "first", "second", "third", etc. are used merely to distinguish one description from another and are not to be construed as indicating or implying relative importance.
Finally, it should be noted that: the above-mentioned embodiments are only specific embodiments of the present invention, which are used for illustrating the technical solutions of the present invention and not for limiting the same, and the protection scope of the present invention is not limited thereto, although the present invention is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art can modify or easily conceive the technical solutions described in the foregoing embodiments or equivalent substitutes for some technical features within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the present invention in its spirit and scope. Are intended to be covered by the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A construction method of an industrial knowledge graph is characterized by comprising the following steps:
acquiring a feature vector of data to be processed;
establishing and initializing an algorithm model, wherein the algorithm model comprises a first deep learning network and a second deep learning network;
generating a first knowledge unit based on the first deep learning network and the feature vector of the data to be processed;
generating a second knowledge unit and weight values of the first knowledge unit and the second knowledge unit based on the second deep learning network and the first knowledge unit;
and generating an industrial knowledge graph containing a directed graph pointed to by the first knowledge unit to the second knowledge unit according to the weight value.
2. The method of claim 1, wherein the obtaining the feature vector of the data to be processed comprises:
establishing a data model, wherein the data model comprises a source end data source, a target end data source and a mapping relation between the source end data source and the target end data source;
acquiring data to be processed from a source end through a source end data source;
and acquiring a target end data source mapped by the data to be processed based on the mapping relation between the source end data source and the target end data source, and acquiring the characteristic vector of the data to be processed.
3. The method of claim 2, wherein initializing the algorithm model comprises:
configuring the mapping of the data to be processed in the first deep learning network and the second deep learning network;
and setting a training algorithm of the first deep learning network and a training algorithm of the second deep learning network.
4. The method of claim 1, wherein generating a first knowledge unit based on the first deep learning network and a feature vector of the data to be processed comprises:
inputting the feature vector of the data to be processed into the first deep learning network to obtain the evaluation level of the data to be processed;
and inputting the evaluation level of the data to be processed into a knowledge base, and enabling the knowledge base to generate a first knowledge unit according to the evaluation level of the data to be processed and the mapping relation between the evaluation level of the data to be processed and the knowledge unit.
5. The method of claim 1, wherein generating weight values for a second knowledge unit and the first and second knowledge units based on the second deep learning network and the first knowledge unit comprises:
inputting the first knowledge unit and the feature vector of the data to be processed into a training layer of the second deep learning network to obtain a second knowledge unit output by the training layer;
inputting the first knowledge unit and the second knowledge unit to a weight value generation layer of the second deep learning network, and generating weight values of the first knowledge unit and the second knowledge unit.
6. An apparatus for building an industrial knowledge graph, comprising:
the data acquisition module is used for acquiring a feature vector of the data to be processed;
the model establishing module is used for establishing and initializing an algorithm model, wherein the algorithm model comprises a first deep learning network and a second deep learning network;
the knowledge graph generating module is used for generating a first knowledge unit based on the first deep learning network and the feature vector of the data to be processed, generating weight values of a second knowledge unit and the first knowledge unit and the second knowledge unit based on the second deep learning network and the first knowledge unit, and generating an industrial knowledge graph containing a directed graph directed to the second knowledge unit by the first knowledge unit according to the weight values.
7. The apparatus according to claim 6, wherein the data obtaining module is specifically configured to include:
establishing a data model, wherein the data model comprises a source end data source, a target end data source and a mapping relation between the source end data source and the target end data source;
acquiring data to be processed from a source end through a source end data source;
and acquiring a target end data source mapped by the data to be processed based on the mapping relation between the source end data source and the target end data source, and acquiring the characteristic vector of the data to be processed.
8. The apparatus of claim 7, wherein the model building module is specifically configured to include:
configuring the index of the data to be processed in the first deep learning network and the second deep learning network;
and setting a training algorithm of the first deep learning network and a training algorithm of the second deep learning network.
9. The apparatus of claim 6, wherein the knowledge-graph generating module is specifically configured to include:
inputting the feature vector of the data to be processed into the first deep learning network to obtain the evaluation level of the data to be processed;
and inputting the evaluation level of the data to be processed into a knowledge base, and enabling the knowledge base to generate a first knowledge unit according to the evaluation level of the data to be processed and the mapping relation between the evaluation level of the data to be processed and the knowledge unit.
10. The apparatus of claim 6, wherein the knowledge-graph generating module is further configured to:
inputting the first knowledge unit and the feature vector of the data to be processed into a training layer of the second deep learning network to obtain a second knowledge unit output by the training layer;
inputting the first knowledge unit and the second knowledge unit to a weight value generation layer of the second deep learning network, and generating weight values of the first knowledge unit and the second knowledge unit.
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 CN109739996A (en) | 2019-05-10 |
| CN109739996B true 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) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110957042B (en) * | 2020-01-17 | 2022-12-27 | 广州慧视医疗科技有限公司 | Method for predicting and simulating eye diseases under different conditions based on domain knowledge migration |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
Family Cites Families (1)
| 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 |
-
2018
- 2018-12-29 CN CN201811643488.4A patent/CN109739996B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
|---|
| 一种知识图谱的排序学习个性化推荐算法;杨晋吉 等;《小型微型计算机系统》;20181130;第39卷(第11期);第2419-2423页 * |
| 我国网络计量学研究的知识扩散可视化分析;韩牧哲 等;《图书情报研究》;20161231(第4期);第82-88页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109739996A (en) | 2019-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11741160B1 (en) | Determining states of key performance indicators derived from machine data | |
| CN109643304B (en) | System and method for displaying trademark inquiry search results in interactive graphic representation | |
| US11017764B1 (en) | Predicting follow-on requests to a natural language request received by a natural language processing system | |
| CN106452858B (en) | A kind of method, apparatus and calculating equipment of the identification network user | |
| JP2011138422A (en) | Device, method and program for detecting behavioral-pattern | |
| CN109144964A (en) | log analysis method and device based on machine learning | |
| CN110928957A (en) | Data clustering method and device | |
| CN110633959A (en) | Method, device, equipment and medium for creating approval task based on graph structure | |
| CN109739996B (en) | Construction method and device of industrial knowledge map | |
| CN114722091A (en) | Data processing method, data processing device, storage medium and processor | |
| CN108563648B (en) | Data display method and device, storage medium and electronic device | |
| CN109961163A (en) | Gender prediction's method, apparatus, storage medium and electronic equipment | |
| CN109388385B (en) | Method and apparatus for application development | |
| CN113055209B (en) | Arranging method and device for edge calculation | |
| CN112989030A (en) | Visualization method and system for similar product relationship | |
| CN115801557A (en) | Fault root cause positioning method and device and readable storage medium | |
| JPWO2011016281A1 (en) | Information processing apparatus and program for Bayesian network structure learning | |
| CN107122283B (en) | Temperature monitoring method and device | |
| CN104796478A (en) | Resource recommending method and device | |
| CN110059272A (en) | A kind of page feature recognition methods and device | |
| CN108846103A (en) | A kind of data query method and device | |
| CN112308167A (en) | Data generation method and device, storage medium and electronic equipment | |
| CN117934663A (en) | Method, system, equipment and medium for generating curve by dragging graphic element in DCS picture | |
| CN115718696A (en) | Source code cryptography misuse detection method and device, electronic equipment and storage medium | |
| CN113420214A (en) | Electronic transaction object recommendation method, device and equipment |
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 |