CN110362660A - A kind of Quality of electronic products automatic testing method of knowledge based map - Google Patents
A kind of Quality of electronic products automatic testing method of knowledge based map Download PDFInfo
- Publication number
- CN110362660A CN110362660A CN201910667154.9A CN201910667154A CN110362660A CN 110362660 A CN110362660 A CN 110362660A CN 201910667154 A CN201910667154 A CN 201910667154A CN 110362660 A CN110362660 A CN 110362660A
- Authority
- CN
- China
- Prior art keywords
- product
- data
- quality
- entity
- knowledge
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Educational Administration (AREA)
- Strategic Management (AREA)
- Computational Linguistics (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention relates to a kind of Quality of electronic products automatic testing methods of knowledge based map, belong to Quality of electronic products detection and knowledge mapping field, comprising: building domain knowledge map and quality determining method are realized.By the related normative document of electronic product and technical specification book, product scope knowledge mapping is constructed;It is detected in terms of four for Quality of electronic products detection process, using the method for Ontology Modeling come the mode layer of managerial knowledge map, then data Layer is constructed on the basis of mode layer, entity and entity relationship are extracted using the method for Bi-LSTM+Attention+CRF from multiple isomeric data sources, data fusion generally uses similarity calculating method, data Layer act actually unit by knowledge store in neo4j chart database, product quality detection is carried out using product standard domain knowledge map, add a front-end module, for directly inquiring quality testing data.
Description
Technical field
The invention belongs to Quality of electronic products detection and knowledge mapping field, the electronics for being related to a kind of knowledge based map is produced
Quality automatic testing method.
Background technique
For electronic product at indispensable part in for people's lives, Quality of electronic products quality is people's selection
Primary to require, to the most of method for using artificial detection of Quality of electronic products detection, testing staff is sampled inspection to product
It tests, personnel is supervised to carry out self-test to production, mutually inspection and first inspection, artificial detection can be such that the development cycle extends, the efficiency of exploitation
It reduces, many drawbacks occurs.In addition to this, a large amount of data can be generated in the detection process, but each detection system is mutual
It is independent, mass data redundancy can be made in this way and data structure is inconsistent, the association between data knowledge can not be established, be unfavorable for matter
The whole management of amount detection systems.Therefore, it is necessary to design a kind of Quality of electronic products that can be realized data correlation to detect automatically
Method solves these problems.
Knowledge mapping is that each entity in the real world is described in the form of structuring, relationship, theoretical, and can utilize can
It depending on changing relationship between technical description entity, from building, analyzes, display, the association between mining data, is semantic net of new generation
It realizes, the use of knowledge mapping is solves the problems, such as that data redundancy provides possibility method.For electronic product fault detection, knowledge
Map now applies considerably less on this, therefore this programme creative utilization knowledge mapping scheme constructs product scope map, realizes
Isomeric data fusion, data sharing analyze to product testing data and then realize product quality detection.
Summary of the invention
In view of this, the purpose of the present invention is to provide a kind of Quality of electronic products of the knowledge based map sides of detection automatically
Method, excessive to solve manually to participate in during quality testing, detection efficiency is low, and testing staff's technical level requires height, each
The few problem of data correlation between detection unit.In order to achieve the above objectives, the invention provides the following technical scheme:
A kind of Quality of electronic products detection method of knowledge based map, comprising the following steps:
S1: based on relevant criterion, electronics field knowledge mapping, base of the map as quality testing are constructed
Plinth;
S2: the title of product to be detected, the attribute of detection and value range are monitored;
S3: product attribute to be detected and the attribute in domain knowledge map are matched, and determine detection target, will be opposite
The attribute value answered is inquired, and according to the matching result of attribute value, determines product quality;
S4: addition front-end module, for directly inquiring quality testing data.
Further, in step S1, comprising the following steps:
For electronic product detection process, relevant criterion derives from national standard, professional standard, company standard and product skill
Art description, specific to Quality of electronic products detection process from product specification book, product appearance examination criteria and
It is detected in terms of product function examination criteria, product reliability examination criteria four, determines the mode layer of knowledge mapping;
Entity recognition is carried out to data source using the method for Recognition with Recurrent Neural Network Bi-LSTM+Attention+CRF, in mould
Data Layer is constructed on the basis of formula layer;
Using the entity disambiguation method based on cluster, the result of Entity recognition is disambiguated, obtains entity unambiguously
Information;
Obtained data include the corresponding entity of electronics field, entity relationship and attribute value, select chart database institute
Data are stated to be stored, that is, the data are imported into chart database neo4j, the relationship between node and node is formed, it is complete
At entire product scope knowledge mapping.
It is different for the attribute specification of electronics field, mobile phone electronic product testing attribute is common in: outer
It sees, function, reliability etc., therefore the entity extracted is related to name of product, color, pixel, memory, operating system, processing
The various aspects such as device, memory;
It determines ternary group object and its relationship, Bi-LSTM+Attention+CRF is utilized for mobile phone electronic product characteristic
Method obtains " ProductName-attribute-name-attribute value " triple.Method particularly includes:
(1) text vector in knowledge base is converted to character vector using word2vec method;
(2) character vector is input in Bi-LSTM network, character vector will be inputted, gone over by preceding to LSTM
Information, obtain following information to LSTM by rear, can simultaneously include according to this contextual information by Bi-LSTM model;
(3) Attention mechanism is added, prominent effective keyword improves recognition efficiency;
(4) by the vector after training, it is input to CRF, makes up the limitation tune of each sequence label after Bi-LSTM training
High discrimination.
Further, step S1 the following steps are included:
S11: Quality of electronic products detection process is directed to from product specification book, product appearance examination criteria, product function
It is detected in terms of energy examination criteria and product reliability examination criteria four, determines the mode layer of knowledge mapping;
S12: carrying out Entity recognition to data source using the method for Recognition with Recurrent Neural Network Bi-LSTM+Attention+CRF,
Data Layer is constructed on the basis of mode layer;
S13: the entity disambiguation method based on cluster is used, the result of Entity recognition is disambiguated, is obtained unambiguously
Entity information;
S14: merging the data in disparate databases source using frames match, that is, ontology alignment method, and then
To the completion of knowledge mapping;
S15: pass through step S11-S14, obtained data include the corresponding entity of electronics field, entity relationship and category
Property value, select chart database described in data stored, that is, the data are imported into chart database neo4j, formed node
Relationship between node completes entire product scope knowledge mapping.
Further, step S11, specifically includes the following steps:
S111: the mode detected according to present Quality of electronic products, from product specification book, product appearance detection mark
Standard and product function examination criteria detect in terms of product reliability examination criteria four, carry out system according to product testing object and build
Mould forms product testing body construction;
S112: the concept and classification relation of clear product testing object, will test object concept is complete and objective shape
At an entirety;
S113: defining may deposit between the knowledge description system, classification and each test object of each test object
Relationship;
S114: the ontology detected by representation of knowledge formal product.
Further, step S12 specifically includes the following steps:
S121: the data of data Layer include the technical specification book of electronic product standard detection document and production,
Belong to unstructured data, by unstructured data using text as input;
S122: using the method migration of word2vec at corresponding character vector, using CBOW model, in CBOW model
In, there is primitive sentence to speculate to obtain target word, that is, passes through known current word wn, utilize the word of front and back 2k, thus it is speculated that existing for current word
Probability P:
P(wn|SUM(wn-k,wn-k+1…,wn-1+k,wn+k)) (1)
To calculate the probability of sentence S using CBOW model, calculation expression is as follows:
T is input text size, and P (S) indicates sentence with (w1,w2,…,wT) occur probability, that is, each word joint
Probability;Then likelihood function is constructed to entire text:
To derive log-likelihood function:
The purpose of CBOW model training is exactly that target likelihood function is made to reach maximum value;
S123: when the character of input obtains corresponding character vector by Word2vec, Bi-LSTM network is input to this
In, character vector will be inputted, obtain past information to LSTM by preceding, obtains following information to LSTM by rear, according to this
It can simultaneously include contextual information by Bi-LSTM model;
S124: in order to improve recognition efficiency, focal point entity, is added Attention mechanism emphatically;
S125: when character vector is by being input to CRF, playing the role of a multi-categorizer after Attention mechanism
Make up the limitation of each word sequence label after Bi-LSTM is trained;
For a list entries Z=(z1,z2,…,zn), wherein n is list entries length, ziIndicate input vector
I-th of word, then the corresponding output sequence label of Z is Y=(y1,y2,…,yn);For in the condition that given sequence Z value is z
Under, the conditional probability that the value on annotated sequence Y is y is p (y | z)
S (z)=∑yexp(∑N,kλktk(yN-1,yN,z,N)+∑N,lμl sl(yN,z,N)) (6)
N=1,2 ..., n, sl() and tk() is characteristic function, μlAnd λkIt is its corresponding weight respectively, S (z) is rule
The generalized factor;In the training stage, optimal sequence is asked to mark with maximal possibility estimation, corresponding likelihood logarithm are as follows: ∑Nlogp(y
| z), highest conditional probability y is obtained by training, list entries is labeled:
Further, step S13 is specifically included:
Using the entity disambiguation method based on cluster, gives entity to be disambiguated and censure set O=O1,O2,…,On, to every
One entity censures O and extracts its feature, is denoted as feature vector, the similarity between computational entity, using the calculation of cluster
Method censures item cluster to entity, so that each of cluster result classification all corresponds on specific target entity, calculates phase
Cosine similarity calculation method is used like degree, calculation formula is as follows:
Above formula indicates similarity size between two n-dimensional vectors A, B, and value shows more similar closer to 1.
Further, step S2 the following steps are included:
S21: it is different for the content of electronics field detection, by taking mobile phone electronic product as an example, determine the product of detection
Type
S22: the attribute for determining product includes: appearance, function, and reliability etc. determines that the type of attribute has ProductName
The various aspects such as title, color, pixel, memory, operating system, processor, memory;
S23: according to every attribute, attribute value is determined.
Further, step S3 the following steps are included:
S31: detecting the practical index obtained according to electronic product in actual production process, and one kind as lane database is non-
Structured data type is passed in product standard knowledge mapping, is converted to chart database type deposit Neo4j chart database
In;
S32: data are stored in map, are carried out retrieval and inquisition using map, are passed through retrieval standard of comparison data and detection
Whether data are consistent, show whether testing result complies with standard, to complete product quality detection.
Further, step S4 is specifically included: one front-end module of addition is given user one basic displayed page, is provided
The data of one basic needs for taking data-interface load mapping diagram data and drawing, then directly from data-interface number of request
According to calling drawing JS completes to draw in the page, can be visually seen the project of inquiry in relation to connection and attribute value, there are neo4j figure numbers
According to the data in library, node and side all have been assigned corresponding key-value pair attribute, when front end page issues http request, by phase
It closes conditional parameter to be transmitted in chart database as inquiry theme, reads respective counts from neo4j database using JavaScript
According to being handled by backstage JavaScript data, then to inquiry, theme corresponding node data are tagged reaches web
Browser is effectively accurately inquired and is shown to product testing object in this way, intuitive to find that whether wrong product quality is complete
At product quality Detection task.
The beneficial effects of the present invention are: the present invention using electronic product normative document and technical specification book as knowing
The standard source for knowing map is limited and the range and content of clear product testing using the method for mode layer ontological construction, then
Entity and entity relationship in normative document and specification are extracted using the method for Bi-LSTM+Attention+CRF, is utilized
The method that similarity determines solves the problems, such as data fusion, the entity of extraction and relationship is deposited into chart database neo4j, most
Domain knowledge map is formed afterwards, carries out product quality detection using electronic product standardization areas knowledge mapping.
When detecting in process of production to each product detection module, the index that directly will test, which is input to, to be known
Know in library, the rule data corresponding with the module in knowledge mapping in knowledge base are compared by inference machine, compare the detection
Attribute corresponding to module and attribute value, if measured data within the scope of profile criteria, otherwise requisite quality is not met
Standard.Using this method, a large amount of human resources can be saved, and four aspects of quality testing can be associated with simultaneously, are mentioned
High detection efficiency finally utilizes JavaScript technology, a front end is arranged, when front end page issues http request, by phase
It closes conditional parameter to be transmitted in chart database as inquiry theme, reads respective counts from neo4j database using JavaScript
According to being handled by backstage JavaScript data, then to inquiry, theme corresponding node data are tagged reaches web
Browser is effectively accurately inquired and is shown to product testing object in this way, intuitive to find that whether wrong product quality is complete
At product quality Detection task.Using the powerful data-handling capacity of knowledge mapping, entire detection system can be connected,
Enterprise is allowed to carry out better data management.
Other advantages, target and feature of the invention will be illustrated in the following description to a certain extent, and
And to a certain extent, based on will be apparent to those skilled in the art to investigating hereafter, Huo Zheke
To be instructed from the practice of the present invention.Target of the invention and other advantages can be realized by following specification and
It obtains.
Detailed description of the invention
To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention is made below in conjunction with attached drawing excellent
The detailed description of choosing, in which:
Fig. 1 is the logical construction schematic diagram for the knowledge mapping that the present invention constructs;
Fig. 2 is that electronic product detection body of the present invention constructs flow chart;
Fig. 3 is the method flow diagram of present invention name Entity recognition;
Fig. 4 is that knowledge mapping constructs flow diagram;
Fig. 5 is the method flow diagram detected to product quality using the product standard knowledge mapping constructed.
Specific embodiment
Illustrate embodiments of the present invention below by way of specific specific example, those skilled in the art can be by this specification
Other advantages and efficacy of the present invention can be easily understood for disclosed content.The present invention can also pass through in addition different specific realities
The mode of applying is embodied or practiced, the various details in this specification can also based on different viewpoints and application, without departing from
Various modifications or alterations are carried out under spirit of the invention.It should be noted that diagram provided in following embodiment is only to show
Meaning mode illustrates basic conception of the invention, and in the absence of conflict, the feature in following embodiment and embodiment can phase
Mutually combination.
Wherein, the drawings are for illustrative purposes only and are merely schematic diagrams, rather than pictorial diagram, should not be understood as to this
The limitation of invention;Embodiment in order to better illustrate the present invention, the certain components of attached drawing have omission, zoom in or out, not
Represent the size of actual product;It will be understood by those skilled in the art that certain known features and its explanation may be omitted and be in attached drawing
It is understood that.
The same or similar label correspond to the same or similar components in the attached drawing of the embodiment of the present invention;It is retouched in of the invention
In stating, it is to be understood that if there is the orientation or positional relationship of the instructions such as term " on ", "lower", "left", "right", "front", "rear"
To be based on the orientation or positional relationship shown in the drawings, be merely for convenience of description of the present invention and simplification of the description, rather than indicate or
It implies that signified device or element must have a particular orientation, be constructed and operated in a specific orientation, therefore is described in attached drawing
The term of positional relationship only for illustration, is not considered as limiting the invention, for the ordinary skill of this field
For personnel, the concrete meaning of above-mentioned term can be understood as the case may be.
Present invention aims at the Quality of electronic products detection methods for designing a kind of knowledge based map, to solve in quality
It is manually participated in excessively in detection process, detection efficiency is low, and testing staff's technical level requires high, data between each detection unit
The few problem of relevance.Technical scheme is as follows:
Firstly, being based on the related normative document of electronic product and technical specification book, product scope knowledge mapping is constructed.
It is detected in terms of four for Quality of electronic products detection process, determines that the mode layer of knowledge mapping, the mode layer of map are true
The normalization of the Upper Concept of electronic product detection is determined.Using the method for Ontology Modeling come the mode layer of managerial knowledge map,
To which by ontology library, for axiom, rule, the tenability of constraint condition is come the entity of specification extraction, entity relationship, entity
Connection between numerous objects such as type and attribute.Then data Layer is constructed on the basis of mode layer, from multiple isomery numbers
Entity is extracted according to the method for using Bi-LSTM+Attention+CRF in source and entity relationship, data fusion generally use phase
Like degree calculation method, determine whether it is the same entity described using the entity similarity and attributes similarity, to carry out
Data fusion.Data Layer act actually unit by knowledge store in neo4j chart database, the knowledge of extraction is formed with " real
The triple of body-relation-entity " or " entity-attribute-attribute value " type can make to know according to the data Layer that mode layer constructs
The knowledge for knowing library redundancy is reduced, and the structure of map is more clear.All data connections finally will be formed into a huge entity
Relational network, i.e. product scope knowledge mapping.
Product quality detection is carried out using product standard domain knowledge map, is produced first according to electronics in actual production process
The practical index that product examine measures out is passed to product standard knowledge graph as a kind of unstructured data types of lane database
In spectrum, chart database type deposit Neo4j is converted to, Neo4j is a kind of novel chart database, can be realized traditional relational number
Mass data is also handled based on graph theory other than the function of analyzing and processing data according to the storage that library is supported.Data are stored in and are schemed
In spectrum, inquiry and reasoning are carried out using map, the rule in knowledge base and the fact in knowledge mapping are compared by inference machine,
It is whether consistent with detection data by comparing normal data, show whether testing result complies with standard, to complete product quality
Detection.
A front-end module is added, gives user one basic displayed page, one is provided and basic adds by data-interface
The data of the needs of mapping diagram data and drawing are carried, then directly call drawing JS complete in the page from data-interface request data
At drawing, the project of inquiry can be can be visually seen in relation to connection and attribute value, more intuitive can thus see detection quality
Whether comply with standard, this is also beneficial to layman and directly inquires data.
Building electronics field knowledge mapping is divided into mode layer and data Layer, mode layer guide data on constitutive logic
The type of layer data and between connection.The logical construction schematic diagram of knowledge mapping is as shown in Figure 1.
In the content object of mode layer major design product quality detection, the side detected according to present Quality of electronic products
Formula, mainly from product specification book, product appearance examination criteria and product function examination criteria, product reliability detection mark
Four aspect detection of standard, carries out system modeling according to product testing object, forms product testing body construction, can specify product inspection
The concept and classification relation of object are surveyed, and the concept that will test object is complete and objectively forms an entirety, then defines
The knowledge description system of each test object, classification, and its between relationship that may be present, finally by representation of knowledge form
The ontology of product testing.Detailed process is as shown in Figure 2.
After mode layer building is completed, data Layer is constructed on this basis.The data of data Layer rely primarily on electronic product
Standard detection document is such as: " YD/T 1539-2006 GSM CDMA reliability engineering requires and test method ", " YS/T
711-2009 mobile phone and digital product shell aluminium and aluminium alloy plate, band ", " GB/T 9298-1998 paint and varnish paint film
Draw lattice experiment ", " GB/T 5170.1-2016 electric and electronic product environmental test equipment method of inspection part 1: general provisions " etc.
The technical specification book of numerous normative documents and production, these data largely belong to unstructured data.To data
Source carries out respective handling, it is necessary first to carry out corresponding Entity recognition for data source.Entity recognition uses Recognition with Recurrent Neural Network
The method of Bi-LSTM+Attention+CRF, whole training process are as shown in Figure 3.Unstructured data mainly with
Text is as input, and in order to which computer can be identified, it is forwarded with the method for word2vec as corresponding character vector, benefit
There is primitive sentence to speculate to obtain target word in CBOW model with CBOW model, that is, passes through known current word wn, utilize front and back 2k
A word, thus it is speculated that probability P existing for current word:
P(wn|SUM(wn-k,wn-k+1…,wn-1+k,wn+k)) (1)
To calculate the probability of sentence S using CBOW model, calculation expression is as follows:
T is input text size, and P (S) indicates sentence with (w1,w2,…,wT) occur probability, that is, each word joint
Probability.Then likelihood function is constructed to entire text:
To derive log-likelihood function:
The purpose of CBOW model training is exactly that target likelihood function is made to reach maximum value.
When the character of input obtains corresponding character vector by Word2vec, it is input in Bi-LSTM network with this.It will be defeated
Enter character vector, obtain past information to LSTM by preceding, obtains following information to LSTM by rear, pass through Bi- according to this
LSTM model can include contextual information simultaneously.In order to improve recognition efficiency, focal point entity, is added emphatically
Attention mechanism;
When character vector is by being input to CRF after Attention mechanism, the purpose for being added CRF layers is in mark sentence
Words, entire text sequence can be labeled from sentence level, it can play the role of a multi-categorizer and make up
The limitation of each word sequence label after Bi-LSTM training.CRF is not only available from sentence level research sequence signature
Each optimal entity tag sequence can also learn the restriction rule of rear face labels automatically in the training process.For one
List entries Z=(z1,z2,…,zn), wherein n is list entries length, ziIndicate i-th of word of input vector, then Z is corresponding
Output sequence label is Y=(y1,y2,…,yn).For under conditions of given sequence Z value is z, taking on annotated sequence Y
The conditional probability that value is y is p (y | z)
S (z)=∑yexp(∑N,kλktk(yN-1,yN,z,N)+∑N,lμlsl(yN,z,N)) (6)
N=1,2 ..., n, sl() and tk() is characteristic function, μlAnd λkIt is its corresponding weight respectively, S (z) is rule
The generalized factor.In the training stage, optimal sequence is asked to mark with maximal possibility estimation, corresponding likelihood logarithm are as follows: ∑Nlogp(y
| z), highest conditional probability y is obtained by training, list entries is labeled:
It can complete also to complete Relation extraction in entire training process to the identification of name entity by this process,
Attribute extraction.Attribute extraction is that the entity attributes information is extracted from text, entity attributes can be regarded as entity
A kind of noun sexual intercourse between attribute, therefore the problem of attribute extraction, is regarded as to one kind of Relation extraction.Relation extraction and
Name Entity recognition equally belongs to a part of information extraction, and a polytypic problem is belonged to from the point of view of essence, utilizes Bi-
LSTM extracts advanced features from word embeding layer, and then Attention mechanism generates a weight vectors, passes through Bi-LSTM's
Output valve is multiplied with weight vectors, so that vocabulary grade feature is merged into the feature of Sentence-level in iteration each time, finally by sentence
The feature vector of grade is classified for relationship.It can be provided to avoid in conventional method dependent on some existing vocabulary using this method
The feature of source, system or manual extraction, reduces computation complexity.
Obtained related entities, but entity have ambiguity to, therefore identify the result is that being difficult to be placed directly on picture library
In, it is necessary to the result of Entity recognition is disambiguated, entity information unambiguously is obtained.Using the entity disambiguation side based on cluster
Method gives entity to be disambiguated and censures set O=O1,O2,…,On, O is censured to each entity and extracts its feature, is indicated
For feature vector, similarity between computational entity can censure item cluster using the algorithm of cluster, so that cluster result to entity
Each of classification all correspond on specific target entity.It calculates similarity and uses cosine similarity calculation method, calculate
Formula is as follows:
Above formula indicates similarity size between two n-dimensional vectors A, B, and value shows more similar closer to 1.
Knowledge fusion: being handled for the data in disparate databases source, in the electronic product knowledge graph of building standard
Time spectrum, data source are mainly field standard and product description, but when carrying out quality testing to product, it will to production
Concrete type electronic product and the data of requirement are merged, because target when product testing is more clear, for four sides
Face, product specification book, product appearance detection, product functionality detection and product reliability detection, so utilizing frame
Method with the alignment of i.e. ontology merge and then obtains the completion of knowledge mapping.
By above-mentioned technology, the corresponding entity of electronics field, entity relationship are obtained, attribute value is examined in the art
It is to be determined when survey according to entity attributes value, so selection chart database carries out data storage, the storage based on graph structure
Mode can directly and accurately reflect knowledge mapping internal structure, be conducive to the inquiry reasoning of knowledge, therefore import data to
In chart database neo4j, the relationship between node and node is formed, completes entire product scope knowledge mapping, detailed process is such as
Shown in Fig. 4.
It is the map of electronic product examination criteria according to the knowledge mapping that above-mentioned steps construct, in production to product testing
When project and index detected, will carry out retrieval and inquisition by map, and realize the product quality detection of knowledge based map,
Flow chart is as shown in Figure 5.
In knowledge mapping, the various aspects type of product testing standard is had been built up, index, attribute value is i.e. normal to be surveyed
The range of examination value or other attributes, in actual product detection process, main whether testing product reaches exactly in terms of four
Standard, therefore the data actually obtained are deposited into knowledge base, then in the way of graph search by knowledge base data and
Data in map carry out true matching, indicate normal if being consistent with the attribute value in map, otherwise just not up to standard, the detection
Method can simultaneously inquire the cubic face data of detection, so as to improve the efficiency and timeliness of detection, realize and produce
Quality detection, if there is non-compliant part, can use map and Data Integration exist when various aspects, which detect, to be completed
It makes inferences together, is conducive to be the discovery that unit occurs in failure, reduce the investigation time for testing staff.
But it is unfavorable for layman there are the content of knowledge mapping in chart database directly to inquire, therefore is added to one
Front-end module gives user one basic displayed page, provides a basic data-interface of taking and loads mapping diagram data and draw
Then the data of the needs of figure are directly called drawing JS to complete to draw in the page, can be can be visually seen from data-interface request data
The project of inquiry more intuitive can thus see whether detection quality complies with standard in relation to connection and attribute value.
There are the data in neo4j chart database, node and side all have been assigned corresponding key-value pair attribute, current end page
Face issue http request when, by correlated condition parameter be transmitted in chart database as inquiry theme, using JavaScript from
Neo4j database reads corresponding data, is handled by backstage JavaScript data, then gives inquiry theme corresponding section
Point data is tagged to reach web browser, and effectively product testing object is accurately inquired and shown in this way, intuitive to send out
The existing whether wrong completion product quality Detection task of product quality.
This programme can be applied in electronics field, idea and method be it is general, for some specific electronics
Product can first construct the standard knowledge map of this field, then in the actual production process, carry out to the data that detection generates
Reasoning inquiry, can detecte whether mass meets related request.For example it in mobile phone quality testing, is said with producing the requirement of production man
Bright and existing industry and national standard are as normal data source, according to the specification of mode layer, extraction available data, opposite
The entity that should be required, attribute value, relationship are put forward using the method for neural network, are stored in database.It is detected in actual production
In the process, it according to the specific Testing index data of each single item, is inputted by front end, retrieval and inquisition is carried out by map, determines detection
Whether index is qualified, to complete quality testing.
Finally, it is stated that the above examples are only used to illustrate the technical scheme of the present invention and are not limiting, although referring to compared with
Good embodiment describes the invention in detail, those skilled in the art should understand that, it can be to skill of the invention
Art scheme is modified or replaced equivalently, and without departing from the objective and range of the technical program, should all be covered in the present invention
Scope of the claims in.
Claims (10)
1. a kind of Quality of electronic products detection method of knowledge based map, it is characterised in that:
S1: based on relevant criterion, electronics field knowledge mapping, basis of the map as quality testing are constructed;
S2: the title of product to be detected, the attribute of detection and value range are monitored;
S3: product attribute to be detected and the attribute in domain knowledge map are matched, and determine detection target, will be corresponding
Attribute value is inquired, and according to the matching result of attribute value, determines product quality;
S4: addition front-end module, for directly inquiring quality testing data.
2. the Quality of electronic products detection method of knowledge based map according to claim 1, it is characterised in that: step S1
In, comprising the following steps:
For electronic product detection process, relevant criterion is from national standard, professional standard, company standard and product technology rule
Lattice specification, specific to Quality of electronic products detection process from product specification book, product appearance examination criteria and product
It is detected in terms of Function detection standard, product reliability examination criteria four, determines the mode layer of knowledge mapping;
Entity recognition is carried out to data source using the method for Recognition with Recurrent Neural Network Bi-LSTM+Attention+CRF, in mode layer
On the basis of construct data Layer;
Using the entity disambiguation method based on cluster, the result of Entity recognition is disambiguated, obtains entity information unambiguously;
Obtained data include the corresponding entity of electronics field, entity relationship and attribute value, select number described in chart database
According to being stored, that is, the data are imported into chart database neo4j, form the relationship between node and node, are completed whole
A product scope knowledge mapping.
3. the Quality of electronic products detection method of knowledge based map according to claim 2, it is characterised in that: determine three
First group object and its relationship obtain " product using Bi-LSTM+Attention+CRF method for mobile phone electronic product characteristic
Name-attribute-name-attribute value " triple, method particularly includes:
(1) text vector in knowledge base is converted to character vector using word2vec method;
(2) character vector is input in Bi-LSTM network, character vector will be inputted, obtain past letter to LSTM by preceding
Breath obtains following information to LSTM by rear, can simultaneously include according to this contextual information by Bi-LSTM model;
(3) Attention mechanism is added, prominent effective keyword improves recognition efficiency;
(4) by the vector after training, it is input to CRF, makes up the limitation height-regulating knowledge of each sequence label after Bi-LSTM training
Not rate.
4. the Quality of electronic products detection method of knowledge based map according to claim 3, it is characterised in that: step S1
The following steps are included:
S11: it is examined for Quality of electronic products detection process from product specification book, product appearance examination criteria, product function
It is detected in terms of mark standard and product reliability examination criteria four, determines the mode layer of knowledge mapping;
S12: Entity recognition is carried out to data source using the method for Recognition with Recurrent Neural Network Bi-LSTM+Attention+CRF, in mould
Data Layer is constructed on the basis of formula layer;
S13: the entity disambiguation method based on cluster is used, the result of Entity recognition is disambiguated, entity unambiguously is obtained
Information;
S14: the data in disparate databases source are merged using frames match, that is, ontology alignment method, and then are known
Know the completion of map;
S15: pass through step S11-S14, obtained data include the corresponding entity of electronics field, entity relationship and attribute
Value, select chart database described in data stored, that is, the data are imported into chart database neo4j, formed node with
Relationship between node completes entire product scope knowledge mapping.
5. the Quality of electronic products detection method of knowledge based map according to claim 4, it is characterised in that: step
S11, specifically includes the following steps:
S111: the mode detected according to present Quality of electronic products, from product specification book, product appearance examination criteria, and
Product function examination criteria detects in terms of product reliability examination criteria four, carries out system modeling, shape according to product testing object
At product testing body construction;
S112: the concept and classification relation of clear product testing object will test the concept of object completely and objectively form one
A entirety;
S113: it defines that may be present between the knowledge description system, classification and each test object of each test object
Relationship;
S114: the ontology detected by representation of knowledge formal product.
6. the Quality of electronic products detection method of knowledge based map according to claim 4, it is characterised in that: step
S12 specifically includes the following steps:
S121: the data of data Layer include the technical specification book of electronic product standard detection document and production, are belonged to
Unstructured data, by unstructured data using text as input;
S122: using the method migration of word2vec at corresponding character vector, have in CBOW model using CBOW model
Primitive sentence speculates to obtain target word, that is, passes through known current word wn, utilize the word of front and back 2k, thus it is speculated that probability P existing for current word:
P(wn|SUM(wn-k,wn-k+1…,wn-1+k,wn+k)) (1)
To calculate the probability of sentence S using CBOW model, calculation expression is as follows:
T is input text size, and P (S) indicates sentence with (w1,w2,…,wT) occur probability, that is, each word joint probability;
Then likelihood function is constructed to entire text:
To derive log-likelihood function:
The purpose of CBOW model training is exactly that target likelihood function is made to reach maximum value;
S123: when the character of input obtains corresponding character vector by Word2vec, being input in Bi-LSTM network with this, will
Character vector is inputted, obtains past information to LSTM by preceding, following information is obtained to LSTM by rear, passes through according to this
Bi-LSTM model can include contextual information simultaneously;
S124: in order to improve recognition efficiency, focal point entity, is added Attention mechanism emphatically;
S125: when character vector is by being input to CRF, playing the role of a multi-categorizer and make up after Attention mechanism
The limitation of each word sequence label after Bi-LSTM training;
For a list entries Z=(z1,z2,…,zn), wherein n is list entries length, ziIndicate i-th of input vector
Word, then the corresponding output sequence label of Z is Y=(y1,y2,…,yn);For marking under conditions of given sequence Z value is z
The conditional probability that value on sequence Y is y is p (y | z)
S (z)=∑yexp(∑N,kλktk(yN-1,yN,z,N)+∑N,lμlsl(yN,z,N)) (6)
N=1,2 ..., n, sl() and tk() is characteristic function, μlAnd λkIt is its corresponding weight respectively, S (z) is standardization
The factor;In the training stage, optimal sequence is asked to mark with maximal possibility estimation, corresponding likelihood logarithm are as follows: ∑Nlogp(y|z),
Highest conditional probability y is obtained by training to be labeled list entries:
7. the Quality of electronic products detection method of knowledge based map according to claim 4, it is characterised in that: step
S13 is specifically included:
Using the entity disambiguation method based on cluster, gives entity to be disambiguated and censure set O=O1,O2,…,On, to each
Entity censures O and extracts its feature, is denoted as feature vector, the similarity between computational entity, using the algorithm pair of cluster
Entity censures item cluster, so that each of cluster result classification all corresponds on specific target entity, calculates similarity
Using cosine similarity calculation method, calculation formula is as follows:
Above formula indicates similarity size between two n-dimensional vectors A, B, and value shows more similar closer to 1.
8. the Quality of electronic products detection method of knowledge based map according to claim 1, it is characterised in that: step S2
The following steps are included:
S21: it is different for the content of electronics field detection, by taking mobile phone electronic product as an example, determine the product type of detection
S22: the attribute for determining product includes: appearance, function, and reliability etc. determines that the type of attribute has name of product, face
The various aspects such as color, pixel, memory, operating system, processor, memory;
S23: according to every attribute, attribute value is determined.
9. the Quality of electronic products detection method of knowledge based map according to claim 1, it is characterised in that: step S3
The following steps are included:
S31: detecting the practical index obtained according to electronic product in actual production process, and one kind as lane database is non-structural
Change data type, be passed in product standard knowledge mapping, is converted in chart database type deposit Neo4j chart database;
S32: data are stored in map, are carried out retrieval and inquisition using map, are passed through retrieval standard of comparison data and detection data
It is whether consistent, show whether testing result complies with standard, to complete product quality detection.
10. the Quality of electronic products detection method of knowledge based map according to claim 1, it is characterised in that: step
S4 is specifically included: one front-end module of addition gives user one basic displayed page, provides one and basic takes data-interface
Then the data of load mapping diagram data and the needs of drawing directly call drawing JS in the page from data-interface request data
Complete to draw, can be visually seen the related connection of project and the attribute value of inquiry, there are the data in neo4j chart database, node with
Side all has been assigned corresponding key-value pair attribute, and when front end page issues http request, correlated condition parameter is transmitted to figure
As inquiry theme in database, corresponding data is read from neo4j database using JavaScript, it is right by backstage
JavaScript data are handled, and then to inquiry, theme corresponding node data are tagged reaches web browser, are had in this way
Effect is accurately inquired and is shown to product testing object, intuitive to find the whether wrong completion product quality detection of product quality
Task.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910667154.9A CN110362660B (en) | 2019-07-23 | 2019-07-23 | Electronic product quality automatic detection method based on knowledge graph |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910667154.9A CN110362660B (en) | 2019-07-23 | 2019-07-23 | Electronic product quality automatic detection method based on knowledge graph |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110362660A true CN110362660A (en) | 2019-10-22 |
CN110362660B CN110362660B (en) | 2023-06-09 |
Family
ID=68219834
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910667154.9A Active CN110362660B (en) | 2019-07-23 | 2019-07-23 | Electronic product quality automatic detection method based on knowledge graph |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110362660B (en) |
Cited By (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110888808A (en) * | 2019-11-16 | 2020-03-17 | 云南湾谷科技有限公司 | Web intelligent test method based on knowledge graph |
CN110990748A (en) * | 2019-12-18 | 2020-04-10 | 成都迪普曼林信息技术有限公司 | National public opinion data acquisition and publishing system |
CN111159421A (en) * | 2019-12-25 | 2020-05-15 | 中国建设银行股份有限公司 | Knowledge graph-based fund query method and device |
CN111160847A (en) * | 2019-12-09 | 2020-05-15 | 中国建设银行股份有限公司 | Method and device for processing flow information |
CN111191851A (en) * | 2020-01-03 | 2020-05-22 | 中国科学院信息工程研究所 | Data center energy efficiency optimization method based on knowledge graph |
CN111506722A (en) * | 2020-06-16 | 2020-08-07 | 平安科技(深圳)有限公司 | Knowledge graph question-answering method, device and equipment based on deep learning technology |
CN111667147A (en) * | 2020-05-11 | 2020-09-15 | 北京方升互联信息技术有限公司 | Professional definition method for detecting service products |
CN111695771A (en) * | 2020-05-07 | 2020-09-22 | 国网安徽省电力有限公司淮南供电公司 | Intelligent electric power material detection, management and control system and method based on Internet of things technology |
CN112084302A (en) * | 2020-08-24 | 2020-12-15 | 江苏易达捷信数字科技有限公司 | Method, system, device and storage medium for detecting list data of cost file |
CN112231418A (en) * | 2020-10-15 | 2021-01-15 | 南方电网数字电网研究院有限公司 | Power standard knowledge graph construction method and device, computer equipment and medium |
CN112364173A (en) * | 2020-10-21 | 2021-02-12 | 中国电子科技网络信息安全有限公司 | IP address mechanism tracing method based on knowledge graph |
CN112381438A (en) * | 2020-11-24 | 2021-02-19 | 北京三维天地科技股份有限公司 | Product quality and standard matching model based on fuzzy similarity algorithm |
CN112541600A (en) * | 2020-12-07 | 2021-03-23 | 上海电科智能系统股份有限公司 | Knowledge graph-based auxiliary maintenance decision method |
CN112559806A (en) * | 2020-12-02 | 2021-03-26 | 广东电网有限责任公司广州供电局 | Partial discharge case library construction method |
CN112598563A (en) * | 2020-12-29 | 2021-04-02 | 中国科学技术大学 | Smart city data construction method based on knowledge graph |
CN112732939A (en) * | 2021-01-15 | 2021-04-30 | 中国科学院空天信息创新研究院 | Spatiotemporal knowledge graph construction method, device, medium and equipment based on GraphDB |
CN112801492A (en) * | 2021-01-22 | 2021-05-14 | 中国平安人寿保险股份有限公司 | Knowledge hierarchy based data quality inspection method and device and computer equipment |
CN112883197A (en) * | 2021-02-08 | 2021-06-01 | 广东电网有限责任公司广州供电局 | Knowledge graph construction method and system for closed switch equipment |
WO2021212682A1 (en) * | 2020-04-21 | 2021-10-28 | 平安国际智慧城市科技股份有限公司 | Knowledge extraction method, apparatus, electronic device, and storage medium |
CN113626574A (en) * | 2021-08-19 | 2021-11-09 | 成都数联云算科技有限公司 | Information query method, system, device and medium |
CN113987212A (en) * | 2021-11-17 | 2022-01-28 | 武汉理工大学 | Knowledge graph construction method for process data in numerical control machining field |
CN114330965A (en) * | 2021-10-15 | 2022-04-12 | 西安交通大学 | Knowledge graph-based QI system construction method for non-API petroleum special pipe |
CN114417015A (en) * | 2022-01-26 | 2022-04-29 | 西南交通大学 | Method for constructing maintainability knowledge graph of high-speed train |
CN115544265A (en) * | 2022-09-13 | 2022-12-30 | 南京航空航天大学 | Bearing fault diagnosis method based on bearing fault knowledge graph |
CN117171367A (en) * | 2023-09-26 | 2023-12-05 | 北京泰策科技有限公司 | Specification detection method for specified attribute values of different database tables |
CN118194214A (en) * | 2024-05-20 | 2024-06-14 | 江西博微新技术有限公司 | Three-dimensional inspection method, system, computer and storage medium for power transmission |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180107917A1 (en) * | 2016-10-19 | 2018-04-19 | Ebay Inc. | Applying a quantitative range for qualitative terms |
CN108447534A (en) * | 2018-05-18 | 2018-08-24 | 灵玖中科软件(北京)有限公司 | A kind of electronic health record data quality management method based on NLP |
CN108647791A (en) * | 2018-03-30 | 2018-10-12 | 中国标准化研究院 | A kind of processing method of multi-source automotive safety information, apparatus and system |
EP3407151A1 (en) * | 2017-05-24 | 2018-11-28 | Tata Consultancy Services Limited | Systems and methods for cognitive control of data acquisition for efficient fault diagnosis |
CN109064318A (en) * | 2018-08-24 | 2018-12-21 | 苏宁消费金融有限公司 | A kind of internet financial risks monitoring system of knowledge based map |
CN109145122A (en) * | 2018-08-02 | 2019-01-04 | 北京仿真中心 | A kind of product know-how map construction and querying method and system |
CN109146611A (en) * | 2018-07-16 | 2019-01-04 | 浙江大学 | A kind of electric business product quality credit index analysis method and system |
CN109378053A (en) * | 2018-11-30 | 2019-02-22 | 安徽影联云享医疗科技有限公司 | A kind of knowledge mapping construction method for medical image |
CN109460010A (en) * | 2018-12-18 | 2019-03-12 | 彩虹无线(北京)新技术有限公司 | The vehicle fault detection method, apparatus and storage medium of knowledge based map |
CN109948911A (en) * | 2019-02-27 | 2019-06-28 | 北京邮电大学 | A kind of appraisal procedure calculating networking products Information Security Risk |
CN110032647A (en) * | 2019-03-12 | 2019-07-19 | 埃睿迪信息技术(北京)有限公司 | Method, apparatus and storage medium based on industrial circle building knowledge mapping |
-
2019
- 2019-07-23 CN CN201910667154.9A patent/CN110362660B/en active Active
Patent Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180107917A1 (en) * | 2016-10-19 | 2018-04-19 | Ebay Inc. | Applying a quantitative range for qualitative terms |
EP3407151A1 (en) * | 2017-05-24 | 2018-11-28 | Tata Consultancy Services Limited | Systems and methods for cognitive control of data acquisition for efficient fault diagnosis |
CN108647791A (en) * | 2018-03-30 | 2018-10-12 | 中国标准化研究院 | A kind of processing method of multi-source automotive safety information, apparatus and system |
CN108447534A (en) * | 2018-05-18 | 2018-08-24 | 灵玖中科软件(北京)有限公司 | A kind of electronic health record data quality management method based on NLP |
CN109146611A (en) * | 2018-07-16 | 2019-01-04 | 浙江大学 | A kind of electric business product quality credit index analysis method and system |
CN109145122A (en) * | 2018-08-02 | 2019-01-04 | 北京仿真中心 | A kind of product know-how map construction and querying method and system |
CN109064318A (en) * | 2018-08-24 | 2018-12-21 | 苏宁消费金融有限公司 | A kind of internet financial risks monitoring system of knowledge based map |
CN109378053A (en) * | 2018-11-30 | 2019-02-22 | 安徽影联云享医疗科技有限公司 | A kind of knowledge mapping construction method for medical image |
CN109460010A (en) * | 2018-12-18 | 2019-03-12 | 彩虹无线(北京)新技术有限公司 | The vehicle fault detection method, apparatus and storage medium of knowledge based map |
CN109948911A (en) * | 2019-02-27 | 2019-06-28 | 北京邮电大学 | A kind of appraisal procedure calculating networking products Information Security Risk |
CN110032647A (en) * | 2019-03-12 | 2019-07-19 | 埃睿迪信息技术(北京)有限公司 | Method, apparatus and storage medium based on industrial circle building knowledge mapping |
Non-Patent Citations (6)
Title |
---|
YACHEN TANG等: "Enhancement of Equipment Mangement Using Knowledge Graph" * |
李容: "基于知识图谱的手机质量检测方法研究与实现", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
李秀芳等: "基于生产设备监测数据的故障诊断仪的开发与应用" * |
杨笑然: "基于知识图谱的医疗专家系统", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
韩智等: "基于知识图谱的高铁动车设备检测系统的本体", 《现代电子技术》 * |
黄丹丹等: "融合attention机制的BI-LSTM-CRF", 《软件》 * |
Cited By (38)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110888808A (en) * | 2019-11-16 | 2020-03-17 | 云南湾谷科技有限公司 | Web intelligent test method based on knowledge graph |
CN111160847A (en) * | 2019-12-09 | 2020-05-15 | 中国建设银行股份有限公司 | Method and device for processing flow information |
CN111160847B (en) * | 2019-12-09 | 2023-08-25 | 中国建设银行股份有限公司 | Method and device for processing flow information |
CN110990748A (en) * | 2019-12-18 | 2020-04-10 | 成都迪普曼林信息技术有限公司 | National public opinion data acquisition and publishing system |
CN111159421A (en) * | 2019-12-25 | 2020-05-15 | 中国建设银行股份有限公司 | Knowledge graph-based fund query method and device |
CN111191851A (en) * | 2020-01-03 | 2020-05-22 | 中国科学院信息工程研究所 | Data center energy efficiency optimization method based on knowledge graph |
CN111191851B (en) * | 2020-01-03 | 2023-06-23 | 中国科学院信息工程研究所 | Knowledge graph-based data center energy efficiency optimization method |
WO2021212682A1 (en) * | 2020-04-21 | 2021-10-28 | 平安国际智慧城市科技股份有限公司 | Knowledge extraction method, apparatus, electronic device, and storage medium |
CN111695771B (en) * | 2020-05-07 | 2024-02-27 | 国网安徽省电力有限公司淮南供电公司 | Electric power material intelligent detection management and control system and method based on Internet of things technology |
CN111695771A (en) * | 2020-05-07 | 2020-09-22 | 国网安徽省电力有限公司淮南供电公司 | Intelligent electric power material detection, management and control system and method based on Internet of things technology |
CN111667147A (en) * | 2020-05-11 | 2020-09-15 | 北京方升互联信息技术有限公司 | Professional definition method for detecting service products |
CN111506722A (en) * | 2020-06-16 | 2020-08-07 | 平安科技(深圳)有限公司 | Knowledge graph question-answering method, device and equipment based on deep learning technology |
CN111506722B (en) * | 2020-06-16 | 2024-03-08 | 平安科技(深圳)有限公司 | Knowledge graph question-answering method, device and equipment based on deep learning technology |
CN112084302B (en) * | 2020-08-24 | 2024-04-30 | 江苏易达捷信数字科技有限公司 | Method, system, device and storage medium for detecting inventory data of cost file |
CN112084302A (en) * | 2020-08-24 | 2020-12-15 | 江苏易达捷信数字科技有限公司 | Method, system, device and storage medium for detecting list data of cost file |
CN112231418A (en) * | 2020-10-15 | 2021-01-15 | 南方电网数字电网研究院有限公司 | Power standard knowledge graph construction method and device, computer equipment and medium |
CN112364173A (en) * | 2020-10-21 | 2021-02-12 | 中国电子科技网络信息安全有限公司 | IP address mechanism tracing method based on knowledge graph |
CN112364173B (en) * | 2020-10-21 | 2022-03-18 | 中国电子科技网络信息安全有限公司 | IP address mechanism tracing method based on knowledge graph |
CN112381438A (en) * | 2020-11-24 | 2021-02-19 | 北京三维天地科技股份有限公司 | Product quality and standard matching model based on fuzzy similarity algorithm |
CN112559806A (en) * | 2020-12-02 | 2021-03-26 | 广东电网有限责任公司广州供电局 | Partial discharge case library construction method |
CN112541600A (en) * | 2020-12-07 | 2021-03-23 | 上海电科智能系统股份有限公司 | Knowledge graph-based auxiliary maintenance decision method |
CN112598563B (en) * | 2020-12-29 | 2023-11-17 | 中国科学技术大学 | Smart city data construction method based on knowledge graph |
CN112598563A (en) * | 2020-12-29 | 2021-04-02 | 中国科学技术大学 | Smart city data construction method based on knowledge graph |
CN112732939A (en) * | 2021-01-15 | 2021-04-30 | 中国科学院空天信息创新研究院 | Spatiotemporal knowledge graph construction method, device, medium and equipment based on GraphDB |
CN112732939B (en) * | 2021-01-15 | 2022-11-29 | 中国科学院空天信息创新研究院 | Spatiotemporal knowledge graph construction method, device, medium and equipment based on GraphDB |
CN112801492A (en) * | 2021-01-22 | 2021-05-14 | 中国平安人寿保险股份有限公司 | Knowledge hierarchy based data quality inspection method and device and computer equipment |
CN112883197A (en) * | 2021-02-08 | 2021-06-01 | 广东电网有限责任公司广州供电局 | Knowledge graph construction method and system for closed switch equipment |
CN112883197B (en) * | 2021-02-08 | 2023-02-07 | 广东电网有限责任公司广州供电局 | Knowledge graph construction method and system for closed switch equipment |
CN113626574B (en) * | 2021-08-19 | 2023-08-29 | 成都数联云算科技有限公司 | Information query method, system and device and medium |
CN113626574A (en) * | 2021-08-19 | 2021-11-09 | 成都数联云算科技有限公司 | Information query method, system, device and medium |
CN114330965A (en) * | 2021-10-15 | 2022-04-12 | 西安交通大学 | Knowledge graph-based QI system construction method for non-API petroleum special pipe |
CN113987212A (en) * | 2021-11-17 | 2022-01-28 | 武汉理工大学 | Knowledge graph construction method for process data in numerical control machining field |
CN114417015A (en) * | 2022-01-26 | 2022-04-29 | 西南交通大学 | Method for constructing maintainability knowledge graph of high-speed train |
CN115544265A (en) * | 2022-09-13 | 2022-12-30 | 南京航空航天大学 | Bearing fault diagnosis method based on bearing fault knowledge graph |
CN117171367A (en) * | 2023-09-26 | 2023-12-05 | 北京泰策科技有限公司 | Specification detection method for specified attribute values of different database tables |
CN117171367B (en) * | 2023-09-26 | 2024-04-12 | 北京泰策科技有限公司 | Specification detection method for specified attribute values of different database tables |
CN118194214A (en) * | 2024-05-20 | 2024-06-14 | 江西博微新技术有限公司 | Three-dimensional inspection method, system, computer and storage medium for power transmission |
CN118194214B (en) * | 2024-05-20 | 2024-07-19 | 江西博微新技术有限公司 | Three-dimensional inspection method, system, computer and storage medium for power transmission |
Also Published As
Publication number | Publication date |
---|---|
CN110362660B (en) | 2023-06-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110362660A (en) | A kind of Quality of electronic products automatic testing method of knowledge based map | |
Tang et al. | Frame-wise cross-modal matching for video moment retrieval | |
CN105760507B (en) | Cross-module state topic relativity modeling method based on deep learning | |
WO2019223793A1 (en) | Patent knowledge base construction method, apparatus, computer device, and storage medium | |
CN102902821B (en) | The image high-level semantics mark of much-talked-about topic Network Based, search method and device | |
CN103544242B (en) | Microblog-oriented emotion entity searching system | |
CN110633366B (en) | Short text classification method, device and storage medium | |
CN105550190A (en) | Knowledge graph-oriented cross-media retrieval system | |
Chen et al. | Integrating information theory and adversarial learning for cross-modal retrieval | |
CN114817673A (en) | Cross-modal retrieval method based on modal relation learning | |
CN110008306A (en) | A kind of data relationship analysis method, device and data service system | |
Miao et al. | A dynamic financial knowledge graph based on reinforcement learning and transfer learning | |
CN112613451B (en) | Modeling method of cross-modal text picture retrieval model | |
CN112948547A (en) | Logging knowledge graph construction query method, device, equipment and storage medium | |
Chen et al. | Exploring science-technology linkages: A deep learning-empowered solution | |
CN113821588A (en) | Text processing method and device, electronic equipment and storage medium | |
CN116186381A (en) | Intelligent retrieval recommendation method and system | |
CN116049381A (en) | Knowledge graph-based specific target visual analysis method and system | |
Wu et al. | A novel topic clustering algorithm based on graph neural network for question topic diversity | |
CN118260717A (en) | Internet low-orbit satellite information mining method, system, device and medium | |
Guo | [Retracted] Intelligent Sports Video Classification Based on Deep Neural Network (DNN) Algorithm and Transfer Learning | |
CN117216617A (en) | Text classification model training method, device, computer equipment and storage medium | |
Deng et al. | Covidia: COVID-19 Interdisciplinary Academic Knowledge Graph | |
Wang et al. | [Retracted] Analysis of Network Information Retrieval Method Based on Metadata Ontology | |
Li et al. | Attention with Long‐Term Interval‐Based Deep Sequential Learning for Recommendation |
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 |