CN106203633A - A kind of Bayesian network construction method and system - Google Patents
A kind of Bayesian network construction method and system Download PDFInfo
- Publication number
- CN106203633A CN106203633A CN201610500561.7A CN201610500561A CN106203633A CN 106203633 A CN106203633 A CN 106203633A CN 201610500561 A CN201610500561 A CN 201610500561A CN 106203633 A CN106203633 A CN 106203633A
- Authority
- CN
- China
- Prior art keywords
- node
- present
- logic sequence
- bayesian network
- module
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/01—Probabilistic graphical models, e.g. probabilistic networks
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Pure & Applied Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Algebra (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention provides construction method and the system of a kind of Bayesian network, method includes: step one, acquisition have the node of labeling property value；Step 2, acquisition logic sequence；Step 3, extraction root node；Step 4, the node in logic sequence is carried out independence checking；Step 5, dig up the roots outside node, successively using the node in logic sequence as present node, extract the node maximum with the present node degree of association father node as present node according to independence the result, and extract and there is no the node of child node as leaf node；Step 6, build Bayesian network with root node, father node and leaf node.System includes: data conversion module, logic order module, root node select module, independence authentication module, father node to select module and Bayesian network to build module.The method and system decrease the trouble being manually entered, and improve intelligent level.
Description
Technical field
The present invention relates to probability net technical field, particularly relate to a kind of Bayesian network construction method and system.
Background technology
Bayesian network, also known as belief network, is the extension of Bayes method, is used for representing probability dependency between variable
Graphical model, what it described is the joint probability distribution deferred to of one group of stochastic variable, and is referred to by a set condition probability
A fixed set condition independence assumption, is that current uncertain knowledge is expressed and one of maximally effective theoretical model in reasoning field.From
After 1988 are proposed by Pearl, have become as the focus studied in recent years.
One Bayesian network model is made up of two parts: network structure and conditional probability table.Network structure is one
Individual directed acyclic graph (DAG), the most corresponding stochastic variable of all nodes in figure, directed edge represents and directly relies on pass between variable
System, embodies the feature of the qualitative aspect of domain knowledge.In directed acyclic graph, given father node, each non-independent of it
Descendant node；Conditional probability table or claim local probability distribution, be with each variable association local probability distribution set, set
In element be the father node of given each variable.
At present, build Bayesian network and generally comprise two kinds of methods:
One, the Bayesian network method of network structure is specified.
The method is determined the structure and parameter of Bayesian network by association area expert based on experience, pattra leaves in early days
This net structure is adopted in this way mostly.The complexity that this mode is suitable for problem domain is the highest, and variable is seldom and relation
Application clearly.But, this method not only needs manually to participate in determining the dependence between each parameter, and new having
Node adds fashionable being also required to and manually participates in, and intelligent level is more weak.
Two, the Bayesian network method of universal architecture study.
The structure of Bayesian network is obtained from a large amount of training data learnings by machine learning algorithm.This method be by
Data-driven, it is particularly suitable for available FIELD Data amount relatively big, and domain knowledge is difficult to situation about grasping completely.But
That the method may study obtain with the Bayesian network of domain knowledge indigestion and explanation, and by data acquisition with
The impact of data set excess kurtosis is relatively big, and presents symmetrical situation, institute in the relatedness of data owing to cause effect relation shows
Can not embody the cause effect relation of nodes, and then usually can obtain the network topology of multiple equivalence, subnetwork is opened up
Flutterring is inverse cause and effect.
Visible, in above two method, the intelligent level of method one is the lowest, and method two is relatively big by data influence, in
Existing two extreme, be all difficult to meet the demand of real world applications.
Summary of the invention
It is an object of the invention to provide construction method and the system of a kind of Bayesian network, to solve the problems referred to above.
The construction method of Bayesian network disclosed in this invention, including:
Step one, unified raw data format, it is thus achieved that there is the node of labeling property value；
Step 2, according to domain knowledge, described node is ranked up, it is thus achieved that logic sequence；
Step 3, extract the highest node that sorts in described logic sequence as root node；
Step 4, the node in described logic sequence is carried out independence checking；
Step 5, successively using the node in described logic sequence as present node, extract according to independence the result
The node maximum with described present node degree of association is as the father node of present node, and extracts the node conduct not having child node
Leaf node；
Step 6, with described root node, father node and leaf node build Bayesian network.
Bayesian network constructing system disclosed in this invention, including:
Data conversion module, described data conversion module is used for unifying raw data format, it is thus achieved that have labeling attribute
The node of value；
Logic order module, described logic order module is connected with described data conversion module, is used for receiving described node,
And by manually according to domain knowledge, described node being ranked up, it is thus achieved that logic sequence；
Root node selects module, and described root node selects module to be connected with described logic order module, be used for receiving described in
Logic sequence, and extract the highest node that sorts in described logic sequence as root node, by described in described logic sequence
Root node is deleted；
Independence authentication module, described independence authentication module selects module to be connected with described root node, is used for receiving institute
State logic sequence, and the node in described logic sequence is carried out independence checking；
Father node selects module, and described father node selects module to be connected with described independence authentication module, for successively will
Node in described logic sequence, as present node, extracts the node maximum with described present node degree of association as working as prosthomere
Point father node, and extract there is no the node of child node as leaf node；
Bayesian network builds module, and described Bayesian network builds module and selects module, father node with root node respectively
Selection module is connected, and is used for receiving described root node, father node and leaf node, and with described root node, father node and leaf
Node builds Bayesian network.
The construction method of Bayesian network disclosed in this invention and system, after early stage unified parameters form, according to neck
Domain knowledge confirms that each internodal logical relation also sorts, it is to avoid the impact that described Bayesian network is brought by data, rear
Phase utilizes the mode of automatization, according to the logical relation of association area, each node is ranked up obtaining root node, and according to
Each internodal degree of association, confirm internodal network structure relation, build Bayesian network with this, decreases and is manually entered
Trouble, improves intelligent level.
Accompanying drawing explanation
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, the present invention will be implemented below
In example or description of the prior art, the required accompanying drawing used is briefly described, it should be apparent that, the accompanying drawing in describing below is only
It is only some embodiments of the present invention, for those skilled in the art, on the premise of not paying creative work, also may be used
To obtain other accompanying drawing according to these accompanying drawings.
Fig. 1 is a kind of Bayesian network construction method schematic flow sheet disclosed in the embodiment of the present invention；
Fig. 2 is the schematic flow sheet of Bayesian network construction method step 5 disclosed in the embodiment of the present invention；
Fig. 3 is a kind of Bayesian network constructing system structural representation disclosed in the embodiment of the present invention；
Fig. 4 is that Bayesian network constructing system father node disclosed in the embodiment of the present invention selects modular structure schematic diagram；
Fig. 5 is another kind of Bayesian network constructing system structural representation disclosed in the embodiment of the present invention.
Detailed description of the invention
For making the purpose of the embodiment of the present invention, technical scheme and advantage clearer, below in conjunction with the embodiment of the present invention
In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is
The a part of embodiment of the present invention rather than whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art
The every other embodiment obtained under not making creative work premise, broadly falls into the scope of protection of the invention.
Embodiment one
Present embodiment discloses the construction method of a kind of Bayesian network, as it is shown in figure 1, the method includes:
Step one, unified raw data format, it is thus achieved that there is the node of labeling property value.Described have labeling attribute
The node of value is the file of unified CSV symbol form, in order to extract attribute, labeling property value.Or have described in,
The node of labeling property value can also can carry out the formatted file of labeling for other to nodal community.
Step 2, according to domain knowledge, described node is ranked up, it is thus achieved that logic sequence.That is, true according to domain knowledge
The direction on fixed each attribute node limit that may be present in Bayesian network, and then represent one or more interrelated
Event in the order of cause and effect.Such as: " alarm classification " field, in causal ordering, higher than " CPU ", because alarm classification is certainly
Determine whether alarm text occurs " CPU " such word.
According to domain knowledge number, order can be complete ordering, it is also possible to be given with hierarchical form.At this
In embodiment, described domain knowledge is the domain knowledge in business network management.
Step 3, extract the highest node that sorts in described logic sequence as root node.The number of described root node is
One or more, then can be made up of one tree in described Bayesian network, it is possibility to have many tree compositions.
Step 4, the node in described logic sequence is carried out independence checking.
Node in described logic sequence carries out independence checking namely the node in described logic sequence is carried out
Pairwisely independent is verified, i.e. in described logic sequence between all of node, to each other relevant of node two-by-two or mutually
Independent checking, and then judge two internodal independences.Described described node is carried out independence checking method be
Card side's proof method, or, covariance coefficient method, or, Pearson is long-pending away from Y-factor method Y.
Such as, card side's proof method, its computing formula is as follows:
Wherein, Observed is the frequency of the combination taking certain value in two features, namely observes frequency,
Expected is expectation frequency, comes from product (independent hypothesis both phases lower that two features individually take the frequency of analog value
Deng).
Step 5, in addition to described root node, successively using the node in described logic sequence as present node, according to independent
Property the result extract the node maximum with the described present node degree of association father node as present node, and extract and there is no son
The node of node is as leaf node.
As in figure 2 it is shown, described step 5 includes:
Step 51, extract in described logic sequence order node after described root node, as present node.Example
As, at logic sequence a, b, c, d, e ... in n, node a is root node, then extract node b at this as present node.
Step 52, described present node is deleted in logic sequence, generate register logic sequence.Such as, in logic sequence
Row a, b, c, d, e ... in n, if node c is present node, then at logic sequence a, b, c, d, e ... in by node c delete, generate
Register logic sequence be a, b, d, e ... n.
Step 53, in described register logic sequence, extract the node maximum with described present node degree of association as working as
The father node of front nodal point.In step 4, the node in described logic sequence is carried out independence checking, it is thus achieved that each two
Internodal degree of association, utilizes the output of step 4 at this, in described register logic sequence, extracts and described present node phase
The node of pass degree maximum is as the father node of present node.
If there is the node that two or more degree of association is maximum, then choose the joint making Bayesian network tree construction height the highest
Point is as the father node of present node.For Bayesian network tree construction, its structure height is the highest, and the relatedness of utilization is just
The most abundant, the situation of distance naive Bayesian is the most remote, and the conditional sampling information contained in tree construction is the most.Choose and make shellfish
The highest node of this network tree structure height of leaf the most just takes full advantage of correlation among nodes as the father node of present node
Information, and mean that the child node quantity of each node is less, be conducive to the simplification calculated.
If the child node that node is present node that described degree of association is maximum, then extract the secondary sport node conduct of degree of association
The father node of present node.Such as, at logic sequence a, b, c, d, e ... in n, during checking node b, maximum with node b degree of association
Be node d, then the father node of node b is node d, i.e. node b is the child node of node d.When verifying node d, with node d
That degree of association is maximum is node b, is node c with the secondary sport node of node d degree of association, then select node c as the father of node d
Node, thus avoid selecting node b and destroying the situation of Bayesian network tree construction.
Step 54, judge whether described present node is last node in logic sequence,
If it is, using there is no child node node as leaf node,
If it is not, then the node that selecting sequence is after described present node is as new present node, go to step 52.
Wherein, if described present node is not last node in logic sequence, then show that logic sequence also has
Node do not process (so-called " process " refers to that its result may find father node to the process of father node finding node,
It is likely to there is no father node) complete, need it is continued with, now need selecting sequence joint after described present node
Point, as new present node, goes to step 52 and continues with it.It is so-called that " selecting sequence is after described present node
Node is as new present node " it is, such as at logic sequence a, b, c, d, e ... in n, if present node is node d, in step
Have determined that the father node of node d in rapid 54, also have the node such as node e, node f to need to verify after node d, now,
After present node d, selecting node e in order is new present node, i.e. after having processed node d, processes node e,
Find the father node of node e, until according to order present in logic sequence, successively by complete for all of node processing.
If described present node is last node in logic sequence, then show joints all in logic sequence
The father node of point all has verified that complete, now, it may appear that do not have the node of child node.Such as, logic sequence a, b, c, d,
E ... in n, the father node of node e is node n, and at logic sequence a, b, c, d, e ... in n, in addition to node e, with other nodes
The maximum node of degree of association be not the most the father node of node e, i.e. other nodes will not be all node e, then node e is leaf joint
Point.
Step 6, with described root node, father node and leaf node build Bayesian network.Determining root node, father
After node and leaf node, i.e. it has been acknowledged that degree of association information mutual between all nodes in described logic sequence, and herein
The father node that exported by above-mentioned steps of father node, described father node not only includes a nodal information, also includes and this joint
The directional information of some interdependent node, accordingly to confirm the node in described Bayesian network and the directional information on limit.Now, described
Root node and leaf node are as the two ends of described Bayesian network, more in addition described father node confirms in described Bayesian network
Node and the directional information on limit, then be enough to build Bayesian network.Finally, then according to prior art corresponding bar is calculated
Part probability tables, completes the structure of described Bayesian network.
The construction method of the Bayesian network disclosed in the present embodiment, after early stage unified parameters form, knows according to field
Know and confirm that each internodal logical relation also sorts, it is to avoid the impact that described Bayesian network is brought by data, in later stage profit
By the mode of automatization, according to the logical relation of association area, each node is ranked up obtaining root node, and according to each joint
Degree of association between point, confirms internodal network structure relation, builds Bayesian network with this, decrease the fiber crops being manually entered
Tired, improve intelligent level.
In the present embodiment, the domain knowledge during described domain knowledge is business network management.Then described in the present embodiment
Bayesian network construction method may be used for the filtration of event in network management system accordingly.That is: the event collected according to webmaster
According to the Bayesian network built, data construct Bayesian network, when corresponding event occurs, judges whether these events are worth
Obtain attendant to pay close attention to.If merited attention, then this event will be distributed to operation maintenance personnel by WorkForm System and carry out manual intervention,
Otherwise this event by screened fall.Through overtesting, the present invention combines operation maintenance personnel micro-judgment, can filter the nothing of more than 75%
Effect event.
Additionally, according to the difference of domain knowledge, the construction method of Bayesian network disclosed in this invention can also be applied
In the middle of the field that other are different, it is not particularly limited at this.
Embodiment two
Present embodiment discloses a kind of Bayesian network constructing system, as it is shown on figure 3, this Bayesian network constructing system bag
Include:
Data conversion module, described data conversion module is used for unifying raw data format, it is thus achieved that have labeling attribute
The node of value；
Logic order module, described logic order module is connected with described data conversion module, is used for receiving described node,
And by manually according to domain knowledge, described node being ranked up, it is thus achieved that logic sequence；
Root node selects module, and described root node selects module to be connected with described logic order module, be used for receiving described in
Logic sequence, and extract the highest node that sorts in described logic sequence as root node, by described in described logic sequence
Root node is deleted；
Independence authentication module, described independence authentication module selects module to be connected with described root node, is used for receiving institute
State logic sequence, and the node in described logic sequence is carried out independence checking；
Father node selects module, and described father node selects module to be connected with described independence authentication module, for successively will
The node dug up the roots in described logic sequence outside node as present node, according to independence the result extract with described currently
The node of node degree of association maximum is as the father node of present node, and extraction does not has the node of child node as leaf node；
Bayesian network builds module, and described Bayesian network builds module and selects module, father node with root node respectively
Selection module is connected, and is used for receiving described root node, father node and leaf node, and with described root node, father node and leaf
Node builds Bayesian network.
As shown in Figure 4, described father node selects module, including:
First present node extraction unit, described first present node extraction unit and described independence authentication module phase
Even, it is used for receiving described logic sequence, extracts order node after described root node in described logic sequence, as currently
Node；
Second present node extraction unit, described second present node extraction unit and described independence authentication module phase
Even, being used for receiving described logic sequence, selecting sequence node after described present node is as new present node；
Register logic sequence generating unit, described register logic sequence generating unit carries with described first present node respectively
Take unit and the second present node extraction unit is connected, for described present node is deleted in logic sequence, generate temporary
Logic sequence；
Father node select unit, described father node select unit respectively with described register logic sequence generating unit and independence
Property authentication module be connected, for receive described present node, register logic sequence and independence authentication module independence checking
As a result, and in described register logic sequence, the node maximum with described present node degree of association is extracted as present node
Father node, if there is the node that two or more degree of association is maximum, then chooses the joint making Bayesian network tree construction height the highest
Point, as the father node of present node, if the child node that node is present node that described degree of association is maximum, then extracts degree of association
Secondary sport node as the father node of present node；
Leaf node select unit, described leaf node select unit respectively with described first present node extraction unit,
Second present node extraction unit selects unit to be connected with father node, is used for receiving described logic sequence, present node and father's joint
Point, and judge that whether described present node is last node of described logic sequence,
If it is, described logic sequence will there is no the node of child node as leaf node,
If it is not, then go to the second present node extraction unit.
As it is shown in figure 5, described Bayesian network constructing system also includes:
Bayesian network output module, described Bayesian network output module builds module phase with described Bayesian network
Even, for receiving and exporting the Bayesian network built.
The constructing system of the Bayesian network disclosed in the present embodiment, after early stage unified parameters form, knows according to field
Know and confirm that each internodal logical relation also sorts, it is to avoid the impact that described Bayesian network is brought by data, in later stage profit
By the mode of automatization, according to the logical relation of association area, each node is ranked up obtaining root node, and according to each joint
Degree of association between point, confirms internodal network structure relation, builds Bayesian network with this, decrease the fiber crops being manually entered
Tired, improve intelligent level.
Above Bayesian network construction method provided by the present invention and system are described in detail.Used herein
Principle and the embodiment of the present invention are set forth by specific case, and the explanation of above example is only intended to help to understand
The method of the present invention and core concept thereof.It should be pointed out that, to those of ordinary skill in the art, without departing from the present invention
On the premise of principle, it is also possible to the present invention is carried out some improvement and modification, these improve and modification also falls into right of the present invention
In the protection domain required.
Claims (9)
1. the construction method of a Bayesian network, it is characterised in that including:
Step one, unified raw data format, it is thus achieved that there is the node of labeling property value；
Step 2, according to domain knowledge, described node is ranked up, it is thus achieved that logic sequence；
Step 3, extract the highest node that sorts in described logic sequence as root node；
Step 4, the node in described logic sequence is carried out independence checking；
Step 5, in addition to described root node, successively using the node in described logic sequence as present node, test according to independence
Card result extracts the node maximum with the described present node degree of association father node as present node, and extraction does not has child node
Node as leaf node；
Step 6, with described root node, father node and leaf node build Bayesian network.
The construction method of Bayesian network the most according to claim 1, it is characterised in that described in there is labeling property value
Node is the file of unified CSV symbol form.
The construction method of Bayesian network the most according to claim 1, it is characterised in that described domain knowledge is enterprise-level net
Domain knowledge in network management.
The construction method of Bayesian network the most according to claim 1, it is characterised in that described described node is carried out independence
Property checking method be card side's proof method, or, covariance coefficient method, or, Pearson amass away from Y-factor method Y.
The construction method of Bayesian network the most according to claim 1, it is characterised in that the number of described root node is one
Or it is multiple.
The construction method of Bayesian network the most according to claim 1, it is characterised in that described step 5 includes:
Step 51, extract in described logic sequence order node after described root node, as present node；
Step 52, described present node is deleted in logic sequence, generate register logic sequence；
Step 53, in described register logic sequence, extract the node maximum with described present node degree of association as working as prosthomere
The father node of point；
If there is the node that two or more degree of association is maximum, then choose the node making Bayesian network tree construction height the highest and make
Father node for present node；
If the child node that node is present node that described degree of association is maximum, then extract the secondary sport node of degree of association as currently
The father node of node；
Step 54, judge whether described present node is last node in logic sequence,
If it is, using there is no child node node as leaf node,
If it is not, then the node that selecting sequence is after described present node is as new present node, and go to step 52.
7. a Bayesian network constructing system, it is characterised in that including:
Data conversion module, described data conversion module is used for unifying raw data format, it is thus achieved that have labeling property value
Node；
Logic order module, described logic order module is connected with described data conversion module, is used for receiving described node, and by
Manually according to domain knowledge, described node is ranked up, it is thus achieved that logic sequence；
Root node selects module, and described root node selects module to be connected with described logic order module, is used for receiving described logic
Sequence, and extract the highest node that sorts in described logic sequence as root node；
Independence authentication module, described independence authentication module and described root node select module to be connected, be used for receiving described in patrol
Collect sequence, and the node in described logic sequence is carried out independence checking；
Father node selects module, and described father node selects module to be connected with described independence authentication module, for successively by described
The node dug up the roots in logic sequence outside node, as present node, extracts and described present node according to independence the result
The node of degree of association maximum is as the father node of present node, and extraction does not has the node of child node as leaf node；
Bayesian network builds module, and described Bayesian network builds module and selects module, father node to select with root node respectively
Module is connected, and is used for receiving described root node, father node and leaf node, and with described root node, father node and leaf node
Build Bayesian network.
Bayesian network constructing system the most according to claim 7, it is characterised in that described father node selects module, including:
First present node extraction unit, described first present node extraction unit is connected with described independence authentication module, uses
In receiving described logic sequence, extract order node after described root node in described logic sequence, as present node；
Second present node extraction unit, described second present node extraction unit is connected with described independence authentication module, uses
In receiving described logic sequence, selecting sequence node after described present node is as new present node；
Register logic sequence generating unit, described register logic sequence generating unit is extracted single with described first present node respectively
Unit is connected with the second present node extraction unit, for being deleted in logic sequence by described present node, generates register logic
Sequence；
Father node selects unit, and described father node selects unit to test with described register logic sequence generating unit and independence respectively
Card module is connected, for receiving the independence the result of described present node, register logic sequence and independence authentication module,
And in described register logic sequence, extract the node maximum with described present node degree of association father's joint as present node
Point, if there is the node that two or more degree of association is maximum, then chooses the node making Bayesian network tree construction height the highest and makees
For the father node of present node, if the child node that node is present node that described degree of association is maximum, then extract the secondary of degree of association
Sport node is as the father node of present node；
Leaf node select unit, described leaf node select unit respectively with described first present node extraction unit, second
Present node extraction unit selects unit to be connected with father node, is used for receiving described logic sequence, present node and father node, and
Judge that whether described present node is last node of described logic sequence,
If it is, described logic sequence will there is no the node of child node as leaf node,
If it is not, then go to the second present node extraction unit.
Bayesian network constructing system the most according to claim 8, it is characterised in that described Bayesian network constructing system is also
Including:
Bayesian network output module, described Bayesian network output module builds module with described Bayesian network and is connected, uses
In receiving and exporting the Bayesian network built.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610500561.7A CN106203633A (en) | 2016-06-29 | 2016-06-29 | A kind of Bayesian network construction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610500561.7A CN106203633A (en) | 2016-06-29 | 2016-06-29 | A kind of Bayesian network construction method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106203633A true CN106203633A (en) | 2016-12-07 |
Family
ID=57463879
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610500561.7A Pending CN106203633A (en) | 2016-06-29 | 2016-06-29 | A kind of Bayesian network construction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106203633A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108614536A (en) * | 2018-06-11 | 2018-10-02 | 云南中烟工业有限责任公司 | A kind of complex network construction method of cigarette primary processing technology key factor |
| CN110362854A (en) * | 2019-05-22 | 2019-10-22 | 北京航天发射技术研究所 | Automatic processing method, the device of a kind of fault tree mathematics library node layout |
| CN110851265A (en) * | 2018-07-25 | 2020-02-28 | 华为技术有限公司 | Data processing method, related equipment and system |
-
2016
- 2016-06-29 CN CN201610500561.7A patent/CN106203633A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108614536A (en) * | 2018-06-11 | 2018-10-02 | 云南中烟工业有限责任公司 | A kind of complex network construction method of cigarette primary processing technology key factor |
| CN108614536B (en) * | 2018-06-11 | 2020-10-27 | 云南中烟工业有限责任公司 | Complex network construction method for key factors of cigarette shred making process |
| CN110851265A (en) * | 2018-07-25 | 2020-02-28 | 华为技术有限公司 | Data processing method, related equipment and system |
| CN110851265B (en) * | 2018-07-25 | 2023-09-08 | 华为云计算技术有限公司 | Data processing method, related equipment and system |
| CN110362854A (en) * | 2019-05-22 | 2019-10-22 | 北京航天发射技术研究所 | Automatic processing method, the device of a kind of fault tree mathematics library node layout |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106326585B (en) | Prediction analysis method and device based on Bayesian Network Inference | |
| Shachter et al. | Decision making using probabilistic inference methods | |
| DE112010006117B4 (en) | Communication management device and data communication method | |
| CN109491816A (en) | Knowledge based engineering method for diagnosing faults | |
| CN109685110B (en) | Training method of image classification network, image classification method and device, and server | |
| CN106650725A (en) | Full convolutional neural network-based candidate text box generation and text detection method | |
| Krzysztofowicz et al. | Fusion of detection probabilities and comparison of multisensor systems | |
| CN106203633A (en) | A kind of Bayesian network construction method and system | |
| CN103902816A (en) | Electrification detection data processing method based on data mining technology | |
| CN108520365A (en) | Education decision system based on big data analysis | |
| CN104506340A (en) | Creation method of decision tree in industrial Ethernet fault diagnosis method | |
| CN106874963A (en) | A kind of Fault Diagnosis Method for Distribution Networks and system based on big data technology | |
| CN108664241A (en) | A method of SysML models are subjected to simulating, verifying | |
| Li et al. | Few-shot image classification via contrastive self-supervised learning | |
| Sourav et al. | Intelligent identification of jute pests based on transfer learning and deep convolutional neural networks | |
| CN109241361A (en) | Data processing method based on block chain | |
| CN109271362A (en) | A kind of data processing system based on block chain | |
| Yu et al. | Constructing traceability between features and requirements for software product line engineering | |
| Preparata et al. | An approach to artificial nonsymbolic cognition | |
| Ali et al. | Machine learning based crop recommendation system for local farmers of pakistan | |
| CN107276807A (en) | A kind of hierarchical network community tree pruning method based on the dynamic tight ness rating in community | |
| Qiwei-Kong et al. | Factor space: a new idea for artificial intelligence based on causal reasoning | |
| CN110135511A (en) | The determination method, apparatus and electronic equipment of discontinuity surface when electric system | |
| CN112818982B (en) | Agricultural pest image detection method based on depth feature autocorrelation activation | |
| CN104537418A (en) | From-bottom-to-top high-dimension-data causal network learning method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20161207 |