CN108345622A - Model retrieval method based on semantic model frame and device - Google Patents
Model retrieval method based on semantic model frame and device Download PDFInfo
- Publication number
- CN108345622A CN108345622A CN201710061310.8A CN201710061310A CN108345622A CN 108345622 A CN108345622 A CN 108345622A CN 201710061310 A CN201710061310 A CN 201710061310A CN 108345622 A CN108345622 A CN 108345622A
- Authority
- CN
- China
- Prior art keywords
- model
- knowledge
- information
- user
- semantic
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2455—Query execution
- G06F16/24552—Database cache management
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/242—Query formulation
- G06F16/243—Natural language query formulation
-
- 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
- G06F16/3332—Query translation
- G06F16/3334—Selection or weighting of terms from queries, including natural language queries
-
- 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
- G06F16/335—Filtering based on additional data, e.g. user or group profiles
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
- G06N5/022—Knowledge engineering; Knowledge acquisition
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention provides the model retrieval methods based on semantic model frame, wherein including：Caching step, the pattern query information of caching and analysis user, and cache relevant knowledge；Query steps inquire the model in caching, index and database, compare the model of the model and the inquiry of user's inquiry in caching, index and database, and be ranked up to correlation model, user is sent to as retrieval result to return to ranking results.Model retrieval method and device provided by the invention based on semantic model frame can carry out correlation model retrieval, fast response time, very useful in the nomination process especially in modeling.The present invention can give query analysis and retrieval result carries out relevant knowledge caching, to ensure the quick-searching of modeling.The present invention can carry out self extension under semantic model frame, i.e., the new knowledge of classification information will not be included into semantic model frame.
Description
Technical field
The present invention relates to technical field of automation in industry, more particularly to the model retrieval method based on semantic model frame and
Device.
Background technology
Semantic model is widely used in describing industrial automation system, such as some are used for system emulation, some are used for
Data and relationship are described.However, device and complex control logic with high number in industrial system, and in semantic model
There are problems that in structure and retrieval：
First, most modeling process needs user in depth to understand very much automated system, also user is required to have
By true system " translation " at the ability of a model.Searching system can be used in modeling to provide domain knowledge
(domain knowledge) and recommend (recommendation), modeling process can be made to become more convenient.
Also, common semantic model search engine is preserved using universal way and retrieval model, for example, RDF jena etc.
Mode.Common semantic model search engine does not have specific aim, such as different processing is not done in terms of domain body, such as simultaneously
It does not do and classifies, also do not consider inquiry reply.Model and data also have no specific aim and put on an equal footing, this can cause very big in model
When the performance of search function and bad.
In addition, common semantic model search engine is only capable of searching for matched model, and relevant model cannot be searched for.
Invention content
First aspect present invention provides the model retrieval method based on semantic model frame, wherein includes the following steps：
Caching step, the pattern query information of caching and analysis user, and cache relevant knowledge；Query steps, inquiry is in caching, index
With the model in database, compare the model of the model and the inquiry of user's inquiry in caching, index and database, and right
Correlation model is ranked up, and user is sent to as retrieval result to return to ranking results.It is provided by the invention to be based on semantic mould
The model retrieval method and device of type frame can carry out correlation model retrieval, fast response time, pushing away especially in modeling
It recommends very useful in process.The present invention can give query analysis and retrieval result carries out relevant knowledge caching, to ensure to model
Quick-searching.The present invention can carry out self extension under semantic model frame, i.e., the new knowledge of classification information will be by
It is included in semantic model frame.
Further, further include following steps after the query steps：Storing step stores the query result.
Further, further include following steps before the caching step：Analytical procedure, analyze semantic model framework and
Knowledge data extracts knowledge and generates knowledge related data.
Further, the analytical procedure further includes following steps：The semantic model frame is analyzed with classification information, is given
The knowledge data assigns classification information；Piece segment information is extracted from the semantic model frame and knowledge data；Calculate segment letter
The probability of breath and classification information.
Further, the caching step further includes following steps：The pattern query information of cache user；Analyze user's
Pattern query information, match stop vocabulary and query word remit the pattern query information classification to user；Replicate the same of high probability
The data of sample type.
Second aspect of the present invention provides the model searching device based on semantic model frame, wherein including：Caching dress
It sets, is used to cache and analyze the pattern query information of user, and cache relevant knowledge；Inquiry unit is used to inquire slow
Deposit, index and database in model, compare the model of user's inquiry and the mould of the inquiry in caching, index and database
Type, and correlation model is ranked up, to return to ranking results user is sent to as retrieval result.It is provided by the invention to be based on
The model retrieval method and device of semantic model frame can carry out correlation model retrieval, and fast response time is especially modeling
In nomination process in it is very useful.The present invention can give query analysis and retrieval result carries out relevant knowledge caching, with true
The quick-searching of health care mould.The present invention can carry out self extension, the i.e. not no new knowledge of classification information under semantic model frame
Knowledge will be included into semantic model frame.
Further, further include：Storage device is used to store the query result.
Further, further include：Analytical equipment is used to analyze semantic model framework and knowledge data, extracts knowledge
And generate knowledge related data.
Further, the analytical equipment further includes：Sort module is used to analyze the semantic mould using classification information
Type frame assigns classification information to the knowledge data；Extraction module is used for from the semantic model frame and knowledge data
Extract piece segment information；Computing module is used to calculate piece segment information and the probability of classification information.
Further, the buffer storage further includes：Buffer storage is used for the pattern query information of cache user；Point
Analysis apparatus, is used to analyze the pattern query information of user, and match stop vocabulary and query word remit the pattern query to user
Information is classified；Reproducing unit is used to replicate the same type of data of high probability.
Description of the drawings
Fig. 1 is the system framework figure of the model index based on semantic model accord to a specific embodiment of that present invention；
Fig. 2 schematically illustrates a semantic model frame SMF；
Fig. 3 is schematically illustrated drives template as the assembly line of knowledge data KD；
Fig. 4 is the caching step S1 of the model retrieval method based on semantic model accord to a specific embodiment of that present invention
Flow chart；
Fig. 5 is the query steps S2 of the model retrieval method based on semantic model accord to a specific embodiment of that present invention
Flow chart；
Fig. 6 is the analytical procedure S0 of the model retrieval method based on semantic model accord to a specific embodiment of that present invention
Flow chart.
Reference sign
100 user's enquiry modules
200 search engines
210 query interface modules
220 cache modules
230 sorting modules
240 spatial caches
241 query caching modules
242 knowledge cache modules
250 index spaces
251 classification vocabulary
252 classified indexes
253 query caching modules
300 analysis modules
400 data spaces
410 model datas
420 segment datas
SMF is based on semantic model frame
KD knowledge datas
MQ pattern query information
254 storage queue modules
SR retrieval results
310 sort modules
312 grouped datas
Model under 314 specific types
320 extraction modules
Vocabulary in 321 knowledge data KD
322 segment informations with classification information
330 computing modules
331 segment informations and its frequency with classification information
Specific implementation mode
Below in conjunction with attached drawing, description of specific embodiments of the present invention.
The present invention provides the model index mechanism based on semantic model frame, are based particularly on the retrieval of correlation model.
Semantic model frame and knowledge data can be analyzed and be stored in search engine, which can be used for retrieving relevant mode
Type.It can be obtained compared with fast-response in the retrieval of correlation model using the present invention in modeling, and especially to modeling process
In recommendation step it is highly effective.
Semantic frame model (SMF, Semantic Model Framework) is the knowledge resource of structuring
(structured knowledge resource).The knowledge includes semantic model standard ISA-95, SSN (Semantic
Sensor Ontology) etc., it is considered as different classification informations.Semantic model frame SMF includes two major parts：Core
The heart (core) and knowledge package (Knowledge Package).Semantic knowledge includes that semantic entity discusses standard (semantic in core
Ontology standards), which depict the common general knowledges about industrial automation system.Knowledge in knowledge package includes tool
There is the model of specific classification information.Fig. 2 schematically illustrates a semantic model frame SMF, as shown, " control system
(Control System ') ", " processing factory (Process Plant) ", " vehicle (Vehicle) " and " assembly line (Assembly
Line) " it is classification, and relevant semantic criteria (such as " ISA-95 " and " ISO-15926 ") and template (such as " engine template "
" driving template ") it is assigned to these classification.In addition, all knowledge data KD (such as " template ", " library
(libraries) " and " sample pattern (sample models) ") can all be assigned to one or more classification, and increase to frame
Frame (framework).Semantic frame model with classification information is for identifying inquiry classification (query
) and searched targets (search target) classification.
Model template (Model templates), library, the sample pattern generated from the project of the past can all be considered to be and know
Know data KD.Wherein, template and library are the semantic model come from experience or standard arrangement, such as the libraries Modelica.Fig. 3 examples
Property show that the assembly line as knowledge data KD drives template, be one from the sample pattern generated from preceding project, example
Such as the sample pattern specially set from the assembly line of Ford Motor.As shown in figure 3, " engine (Motor) ", " gear-box
(Gearbox) ", " friction pulley (Roller) ", " shock sensor (Vibration Sensor) " and " displacement sensor
(Displacement Sensor) " is building blocks, and the connection between the line expression between above-mentioned building blocks.For example, " starting
Connection relation of the machine " between " vibrating sensor " and " friction pulley " and " displacement sensor " is connection, " gear-box " difference
Connection relation between " engine " and " friction pulley " is driving.Knowledge data KD is described with unified format, such as
RDF, Modelica language.
Fig. 1 is the system framework figure of the model index based on semantic model accord to a specific embodiment of that present invention,
Including user's enquiry module 100, search engine 200, analysis module 300 and data space 400.Wherein, user's enquiry module 100
For sending Query Information to search engine, and receive query result.Typically, Query Information is matched with inquiry mesh including one
Target model, optionally it include some object modules classification information.Search engine 200 completes caching, index and storage
Search result is returned into user's enquiry module 100 later Deng some row operations.Analysis module 300 is for analyzing semantic model frame
Frame SMF and knowledge data KD extracts knowledge and generates knowledge related data, used in retrieving for search engine 200.Number
According to the data of 400 stored fragments information and model of space, the data of above two type all include corresponding classification information and general
Rate, the data in data space 400 can be found by the location information (position information) stored in index.
Query interface module 210 is looked into as the bridge between user's enquiry module 100 and the nucleus module of search engine 200, reception
It askes information and Query Information is simultaneously sent to cache module 220 and sorting module 230, and receive retrieval result SR and be sent to user and look into
Ask module 100.
Referring to Fig. 1, the model retrieval method provided by the invention based on semantic model frame SMF includes the following steps：
Caching step S1, the pattern query information MQ of caching and analysis user is first carried out, and caches relevant knowledge.Into one
Step ground, as shown in Figure 1 and Figure 4, the caching step further includes following steps：The cache user in query caching module 243
Pattern query information MQ；The pattern query information MQ of user, match stop vocabulary and inquiry are analyzed in query caching module 241
Vocabulary carrys out the pattern query information MQ classification to user；The same type of data of high probability are replicated from hot knowledge index 253
Into knowledge cache module 242.For example, as the pattern query information MQ " generator drive gear-box " for receiving a user, point
Category information is then locked in " assembly line ".As shown in the table, then there is the segment of the sequence first list of " assembly line " of high probability
Information and model can be buffered in knowledge cache module 242.
Table 1 has the piece segment information and model of high probability
| From | To | connection | classification | count | FromProb | ToProb | ConnProb | ClassProb |
| Engine | Gear-box | Driving | Assembly line | 5 | 0.85 | 0.75 | 0.41 | 1.0 |
| Gear-box | Friction pulley | Driving | Assembly line | 4 | 0.83 | 0.63 | 0.58 | 0.94 |
As shown above, " from " and " to " indicate segment in building blocks and connection direction, schematically illustrate as
The connection relation between building blocks representated by the arrow direction of Fig. 3, wherein " connection " indicates connection relation.
" classification " presentation class, " count " indicate the frequency that the segment occurs.“FromProb”、“ToProb”、
" ConnProb " and " ClassProb " is the calculating of each element conditional probability (conditional probabilities).
Wherein, spatial cache 240 as shown in Figure 1 includes query caching module 241 and knowledge cache module 242, wherein
Query caching in a particular time range is considered as the queue in query caching module 241, can be buffered module
220 analyses simultaneously determine the classification of inquiry by comparing vocabulary and classification vocabulary 251 is inquired.Wherein, for knowledge cache module
242, model and segment data 420 of the inquiry with high probability can be copied to from hot knowledge index 253 in same classification knows
Know in cache module 242.
Wherein, index space 250 as shown in Figure 1 includes classification vocabulary 251, classified index 252 and hot knowledge index
253, it is saved in analysis module 300, and read by sorting module 230 and cache module 220.Wherein, classification vocabulary 251 is used
In determine one inquiry or model classification.Classified index 252 is segment data 420 and model data 410 depositing in classification
Storage space is set, and is used to distribute the specific data in classification.Hot knowledge index 253 is the segment information and model for having high probability
It replicates, by classifying come tissue, and for reducing range of search.
Wherein, data space 400 as shown in Figure 1 is used for memory segment data 420 and model data 410, two types
Data all include corresponding classification and probability, above-mentioned data can be found by location information in the index.
Segment data 420 includes that connection between building blocks, building blocks and a building blocks are connected to the general of another building blocks
Rate.Segment data 420 further includes the probability of the building blocks and connection in different classifications.It is all include semantic model frame SMF and
Knowledge data KD can be considered in extracting segment data, some mathematical statistics algorithms (mathematical statistic
Algorithms semantic model frame SMF and knowledge data KD) can be calculated and expressed by a simple form.For example, such as
Shown in Fig. 3, in one " assembly line driving template ", " engine ", " gear-box ", " friction pulley ", " vibrating sensor " and " shifting
Level sensor " is all building blocks, and the connection relation between any two building blocks by the arrow and arrow that illustrate " connection ",
" driving " indicates.The model of minimum unit includes the connection relation between two building blocks and two building blocks, is from shown in Fig. 3
It is extracted in input template, is considered as a segment information.
2 segment Examples of information of table
| From | To | connection | classification | count | FromProb | ToProb | ConnProb | ClassProb |
| Engine | Gear-box | Driving | Assembly line | 5 | 0.85 | 0.75 | 0.41 | 1.0 |
| Engine | Vibrating sensor | Connection | Assembly line | 6 | 0.76 | 0.66 | 0.62 | 0.96 |
| Gear-box | Friction pulley | Driving | Assembly line | 4 | 0.83 | 0.63 | 0.58 | 0.94 |
| Friction pulley | Displacement sensor | Connection | Assembly line | 2 | 0.64 | 0.88 | 0.49 | 1.0 |
Upper table is the example of a piece segment information, and as shown above, " from " and " to " indicates building blocks and connection in segment
Direction, schematically illustrate the connection relation between the building blocks as representated by the arrow direction of Fig. 3, wherein
" connection " indicates connection relation." classification " presentation class, " count " indicate the frequency that the segment occurs
Rate." FromProb ", " ToProb ", " ConnProb " and " ClassProb " is each element conditional probability (conditional
Probabilities calculating).
Wherein, conditional probability P (B | A) indicates the probability of the B when A has occurred and that, wherein A and B indicates different building blocks.Cause
This, we can obtain：
FromProb=P (motor-driven gear case | engine)=0.85；
ToProb=P (motor-driven gear case | gear-box)=0.75；
ConnProb=P (motor-driven gear case | driving)=0.41；
ClassProb=P (assembly line | motor-driven gear case)=1.0
Model data 410 includes model and query rate, and there are one independent memory spaces for each model tool.We can profit
It is stored with common semantic model search engine, such as Jena.The position of model records in classified index 252.
Then execute query steps S2, the model in query caching, index and database, compare user's inquiry model and
Model in the caching, index and database, and correlation model is ranked up, to return to ranking results as retrieval result
It is sent to user.
Specifically, as shown in Figure 1 and Figure 5, S21, pattern query information MQ quilts in knowledge cache module 242 is first carried out
It receives, measure (measurement) and sequence (ranking), and generate a correlation model list (A relative
model list).If the quantity of correlation model is more than predetermined threshold in correlation model list (i.e. result is good enough), then can
Execute following step S24, it is on the contrary then continue to execute step S22.In step S22, survey is continued to execute in hot knowledge index 253
Amount and sequence, also will produce a correlation model list, make a reservation for if the quantity of correlation model compares in correlation model list
Threshold value is more (i.e. result is good enough), then can execute following step S24, on the contrary then continue to execute step S23.In step 23,
Measurement and sequence are continued to execute in classified index 252 and data space 400, also will produce a correlation model list.Finally
Step S24 is executed, ranking results will be placed on storage queue module (saving queue) 254 and knowledge cache module simultaneously
In 242, query rate (query rate) meeting quilt in hot knowledge index 253 and data space 400 of the model and piece segment information
It preserves again.Hot knowledge index 253 can be sorted result and rearrange, for example, do not have in the hot knowledge index 253 model and
Piece segment information, but its query rate is higher than predetermined threshold, and the model and piece segment information will be then copied in hot knowledge index 253.
For example, Query Information Q includes a model with some building blocks and connection as shown in Figure 3, building blocks include " generator ", " tooth
Roller box " and " friction pulley ", specifically " generator " connect and drive " gear-box ", " gear-box " connect simultaneously drives " friction pulley ",
And query result includes relevant segment information and model as shown in the table.
3 correlation model of table and piece segment information
| From | To | connection | classification | count | FromProb | ToProb | ConnProb | ClassProb |
| Engine | Vibrating sensor | Connection | Assembly line | 6 | 0.76 | 0.66 | 0.62 | 0.96 |
| Friction pulley | Displacement sensor | Connection | Assembly line | 2 | 0.64 | 0.88 | 0.49 | 1.0 |
As shown above, " from " and " to " indicate segment in building blocks and connection direction, schematically illustrate as
The connection relation between building blocks representated by the arrow direction of Fig. 3, wherein " connection " indicates connection relation.
" classification " presentation class, " count " indicate the frequency that the segment occurs.“FromProb”、“ToProb”、
" ConnProb " and " ClassProb " is the calculating of each element conditional probability (conditional probabilities).
Specifically the pattern query information MQ principles measured and process will be illustrated below.For segment information measurement,
It, should if than one predetermined threshold of searching probability (combing probabilities) of interrogation model and piece segment information is big
Segment information is considered as retrieval result SR.For model measurement, if interrogation model is included in a model data 410, and
And the building blocks quantitative proportion in interrogation model than big in model data 410 and is more than a predetermined threshold, then pattern number
It can be considered as retrieval result SR according to 410.If interrogation model is not included in a model data 410, and in building blocks class
Type accounts for that than one predetermined threshold of the interrogation model in a model data 410 is big, then the model data 410 is considered as retrieval result
SR。
Optionally, further include analytical procedure S0 before the caching step S1, analyze semanteme model framework SMF and knowledge
Data KD extracts knowledge and generates knowledge related data.Specifically, the analytical procedure S0 further includes following steps：With classification
Semantic model frame SMF described in information analysis assigns classification information to the knowledge data KD；From the semantic model frame
SMF and knowledge data KD extract piece segment information, and will be stored in memory space；Calculate the general of piece segment information and classification information
Rate, and be stored in memory space.Wherein, the data that step S0 is generated by analysis will be stored in memory space, be stored
Related data (reference data) can also be returned to analytical procedure S0 by space.
Specifically, as shown in fig. 6, when semantic model frame SMF is embedded into analysis module 300, the classification in analysis module
Module 310 can preserve grouped data 312 in classified index 252, and the model 314 under specific type can be saved to mould
Type data 410.When knowledge data KD is embedded into analysis module 300, sort module 310 can have classification vocabulary 251 by checking
Knowledge data KD in vocabulary 321 determine classification belonging to knowledge data KD.Extraction module 320 from input model for carrying
It takes vocabulary and is stored in classification vocabulary 251, and generate the segment information 322 with classification information.Computing module 330 is based on
The frequency of point counting segment information, and by with classification information segment information and its frequency 331 is stored in segment data 420 and heat is known
Know in index 253.For example, when our Input knowledge data are " assembly line driving template ", output information can be as shown in Figure 3
Knowledge index comprising the segment information such as " motor-driven gear case " and the model such as " driving template ".
Optionally, further include storing step S3 after the query steps S2, that is, store the query result.
Second aspect of the present invention additionally provides the model searching device based on semantic model frame, wherein including：Caching dress
It sets, is used to cache and analyze the pattern query information of user, and cache relevant knowledge；Inquiry unit is used to inquire slow
Deposit, index and database in model, compare the model of user's inquiry and the mould of the inquiry in caching, index and database
Type, and correlation model is ranked up, to return to ranking results user is sent to as retrieval result.
Further, the model searching device based on semantic model frame further includes storage device, is used to store
The query result.
Further, the model searching device based on semantic model frame further includes analytical equipment, is used to analyze
Semantic model frame and knowledge data extract knowledge and generate knowledge related data.
Further, the analytical equipment further includes：
Sort module 310 is used to analyze the semantic model frame using classification information, be assigned to the knowledge data
Classification information；
Extraction module 320 is used to extract piece segment information from the semantic model frame and knowledge data；
Computing module 330 is used to calculate piece segment information and the probability of classification information.
Further, the buffer storage further includes：
Buffer storage is used for the pattern query information of cache user；
Analytical equipment, is used to analyze the pattern query information of user, and match stop vocabulary and query word are remitted to user
Pattern query information classification；
Reproducing unit is used to replicate the same type of data of high probability.
The function of above-mentioned apparatus and module has been described above in the model retrieval method based on semantic model frame
It is described in detail, for simplicity, repeats no more.
Model retrieval method and device provided by the invention based on semantic model frame can carry out correlation model retrieval,
Its fast response time, it is very useful in the nomination process especially in modeling.The present invention can give query analysis and retrieval is tied
Fruit carries out relevant knowledge caching, to ensure the quick-searching of modeling.The present invention can carry out self under semantic model frame and expand
Exhibition, i.e., the new knowledge of classification information will not be included into semantic model frame.
Although present disclosure is discussed in detail by above preferred embodiment, but it should be appreciated that above-mentioned
Description is not considered as limitation of the present invention.After those skilled in the art have read the above, for the present invention's
A variety of modifications and substitutions all will be apparent.Therefore, protection scope of the present invention should be limited to the appended claims.
In addition, any reference signs in the claims should not be construed as limiting the involved claims；One word of " comprising " is not excluded for
Unlisted device or step in other claims or specification；The words such as " first ", " second " are only used to indicate names, and
It does not represent any particular order.
Claims (10)
1. the model retrieval method based on semantic model frame, wherein include the following steps：
Caching step, the pattern query information of caching and analysis user, and cache relevant knowledge；
Query steps inquire the model in caching, index and database, and the model and the inquiry for comparing user's inquiry are delaying
Deposit, index and database in model, and correlation model is ranked up, is sent to using returning to ranking results as retrieval result
User.
2. the model retrieval method according to claim 1 based on semantic model frame, which is characterized in that in the inquiry
Further include following steps after step：
Storing step stores the query result.
3. the model retrieval method according to claim 1 based on semantic model frame, which is characterized in that in the caching
Further include following steps before step：
Analytical procedure analyzes semantic model framework and knowledge data, extracts knowledge and generates knowledge related data.
4. the model retrieval method according to claim 3 based on semantic model frame, which is characterized in that the analysis step
Rapid further includes following steps：
The semantic model frame is analyzed with classification information, classification information is assigned to the knowledge data；
Piece segment information is extracted from the semantic model frame and knowledge data；
Calculate the probability of piece segment information and classification information.
5. the model retrieval method according to claim 1 based on semantic model frame, which is characterized in that the caching step
Rapid further includes following steps：
The pattern query information of cache user；
The pattern query information of user is analyzed, match stop vocabulary and query word remit the pattern query information classification to user；
Replicate the same type of data of high probability.
6. the model searching device based on semantic model frame, wherein including：
Buffer storage, is used to cache and analyze the pattern query information of user, and caches relevant knowledge；
Inquiry unit is used to inquire the model in caching, index and database, compares the model of user's inquiry and described looks into
The model in caching, index and database is ask, and correlation model is ranked up, to return to ranking results as retrieval result
It is sent to user.
7. the model searching device according to claim 6 based on semantic model frame, which is characterized in that it further includes：
Storage device is used to store the query result.
8. the model searching device according to claim 6 based on semantic model frame, which is characterized in that it further includes：
Analytical equipment is used to analyze semantic model framework and knowledge data, extracts knowledge and generates knowledge related data.
9. the model searching device according to claim 8 based on semantic model frame, which is characterized in that the analysis dress
It sets and further includes：
Sort module (310) is used to analyze the semantic model frame using classification information, assigns and divides to the knowledge data
Category information；
Extraction module (320) is used to extract piece segment information from the semantic model frame and knowledge data；
Computing module (330), is used to calculate piece segment information and the probability of classification information.
10. the model searching device according to claim 1 based on semantic model frame, which is characterized in that the caching
Device further includes：
Buffer storage is used for the pattern query information of cache user；
Analytical equipment, is used to analyze the pattern query information of user, and match stop vocabulary and query word remit the mould to user
Type Query Information is classified；
Reproducing unit is used to replicate the same type of data of high probability.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710061310.8A CN108345622A (en) | 2017-01-25 | 2017-01-25 | Model retrieval method based on semantic model frame and device |
| PCT/EP2018/051839 WO2018138205A1 (en) | 2017-01-25 | 2018-01-25 | Model search method and device based on semantic model framework |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710061310.8A CN108345622A (en) | 2017-01-25 | 2017-01-25 | Model retrieval method based on semantic model frame and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108345622A true CN108345622A (en) | 2018-07-31 |
Family
ID=61157172
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710061310.8A Pending CN108345622A (en) | 2017-01-25 | 2017-01-25 | Model retrieval method based on semantic model frame and device |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN108345622A (en) |
| WO (1) | WO2018138205A1 (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020124377A1 (en) * | 2018-12-18 | 2020-06-25 | 西门子股份公司 | Modeling knowledge acquisition method and apparatus, and system |
| CN113015942A (en) * | 2018-12-28 | 2021-06-22 | 西门子股份公司 | Data analysis method, device and system |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115577526B (en) * | 2022-10-09 | 2023-05-26 | 深圳大学 | Customized building information model construction method for pedestrian traffic modeling |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102999563A (en) * | 2012-11-01 | 2013-03-27 | 无锡成电科大科技发展有限公司 | Network resource semantic retrieval method and system based on resource description framework |
| CN104298784A (en) * | 2014-11-07 | 2015-01-21 | 巨轮股份有限公司 | Die design implicit knowledge acquiring and searching method |
| CN106250371A (en) * | 2008-08-29 | 2016-12-21 | 启创互联公司 | For utilizing the definition of existing territory to carry out the system and method that semantic concept definition and semantic concept relation is comprehensive |
| US20170011092A1 (en) * | 2015-07-10 | 2017-01-12 | Trendkite Inc. | Systems and methods for the creation, update and use of models in finding and analyzing content |
-
2017
- 2017-01-25 CN CN201710061310.8A patent/CN108345622A/en active Pending
-
2018
- 2018-01-25 WO PCT/EP2018/051839 patent/WO2018138205A1/en active Application Filing
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106250371A (en) * | 2008-08-29 | 2016-12-21 | 启创互联公司 | For utilizing the definition of existing territory to carry out the system and method that semantic concept definition and semantic concept relation is comprehensive |
| CN102999563A (en) * | 2012-11-01 | 2013-03-27 | 无锡成电科大科技发展有限公司 | Network resource semantic retrieval method and system based on resource description framework |
| CN104298784A (en) * | 2014-11-07 | 2015-01-21 | 巨轮股份有限公司 | Die design implicit knowledge acquiring and searching method |
| US20170011092A1 (en) * | 2015-07-10 | 2017-01-12 | Trendkite Inc. | Systems and methods for the creation, update and use of models in finding and analyzing content |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020124377A1 (en) * | 2018-12-18 | 2020-06-25 | 西门子股份公司 | Modeling knowledge acquisition method and apparatus, and system |
| CN113168283A (en) * | 2018-12-18 | 2021-07-23 | 西门子股份公司 | Knowledge acquisition method, device and system for modeling |
| CN113015942A (en) * | 2018-12-28 | 2021-06-22 | 西门子股份公司 | Data analysis method, device and system |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2018138205A1 (en) | 2018-08-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108280114B (en) | Deep learning-based user literature reading interest analysis method | |
| CN109240901B (en) | Performance analysis method, performance analysis device, storage medium, and electronic apparatus | |
| KR101027864B1 (en) | Machine-learned approach to determining document relevance for search over large electronic collections of documents | |
| US20060004753A1 (en) | System and method for document analysis, processing and information extraction | |
| US11182364B2 (en) | Data analysis support apparatus and data analysis support method | |
| EP1860578A1 (en) | System for analyzing patents | |
| Bayomie et al. | A probabilistic approach to event-case correlation for process mining | |
| CN1983255A (en) | Internet searching method | |
| KR102249466B1 (en) | Data catalog providing method and system for providing recommendation information using artificial intelligence recommendation model | |
| CN110706015B (en) | Feature selection method for advertisement click rate prediction | |
| CN108345622A (en) | Model retrieval method based on semantic model frame and device | |
| CN103425748B (en) | A kind of document resources advise the method for digging and device of word | |
| Zhang et al. | Analysis and research on library user behavior based on apriori algorithm | |
| Rubei et al. | A lightweight approach for the automated classification and clustering of metamodels | |
| CN110069558A (en) | Data analysing method and terminal device based on deep learning | |
| Banaei et al. | Web page rank estimation in search engine based on SEO parameters using machine learning techniques | |
| Keerthana et al. | Accurate prediction of fake job offers using machine learning | |
| CN116894495A (en) | Method, computer readable medium and system for enhancing machine learning pipeline with annotations | |
| Joppi et al. | POP: mining POtential Performance of new fashion products via webly cross-modal query expansion | |
| CN115328945A (en) | Data asset retrieval method, electronic device and computer-readable storage medium | |
| JP7292235B2 (en) | Analysis support device and analysis support method | |
| CN114328844A (en) | Text data set management method, device, equipment and storage medium | |
| Shang et al. | Alpine meadow: A system for interactive automl | |
| Huang et al. | Information Retrieval from Internet Data Using Textual Mining and Statistical Models for Stock Market Price Movement Prediction | |
| Elsid et al. | Data Mining: Classification Techniques of Students’ Database A Case Study of the Nile Valley University, North Sudan |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180731 |