CN106651317A - Method and device for judging business process correlation - Google Patents
Method and device for judging business process correlation Download PDFInfo
- Publication number
- CN106651317A CN106651317A CN201611231847.6A CN201611231847A CN106651317A CN 106651317 A CN106651317 A CN 106651317A CN 201611231847 A CN201611231847 A CN 201611231847A CN 106651317 A CN106651317 A CN 106651317A
- Authority
- CN
- China
- Prior art keywords
- process2
- process1
- relation
- flow process
- flow
- 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
- 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
- G06Q10/10—Office automation; Time management
- G06Q10/103—Workflow collaboration or project management
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Economics (AREA)
- Marketing (AREA)
- Data Mining & Analysis (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a method and a device for judging a business process correlation, and relates to the field of information processing. The method comprises the steps of determining a process sample, and determining a business weight characteristic for measuring the correlation among processes, wherein the business weight characteristic includes a structure characteristic and a behavior relation characteristic; generating a classified training relation vector based on an association characteristic value and an association target value, and carrying out classified training to obtain a classification model; and judging the correlation between new business processes and the business processes based on the classification model. According to the method and the device disclosed by the invention, two dimensionalities, namely, a structure and a behavior, are integrated, the classified training is carried out based on the association characteristic value and the association target value to generate the classification model, the correlation between the new business processes and the business processes is judged via the classification model, the business processes in a process library can be divided into new business process correlated and uncorrelated processes, and preprocessing work is carried out via similarity retrieval on the processes, so the accuracy rate and the efficiency of the similarity retrieval and the correlation judgment can be improved.
Description
Technical field
The present invention relates to technical field of information processing, more particularly to a kind of method of discrimination and dress of operation flow correlation
Put.
Background technology
Operation flow is the critical asset of company, is played the role of to the operation and management of company important.As company advises
The continuous expansion of mould, management are becoming better and approaching perfection day by day, and increasing procedural model is created, the procedural model number stored in flow process storehouse
According to increasing.For example, telecommunications industry possesses substantial amounts of procedural model, including inquiry service procedure, recharging and paying flow process, business
Handle flow process etc..These procedural models can provide competitive advantage for enterprise, but while also to the management band of business process model
Carry out certain difficulty.In order to adapt to new demand, the process modeling personnel of company will carry out weight to the handling process of certain business
New design, before new flow process is added to into flow process storehouse, in order to avoid redundancy, need first the related procedure in flow process storehouse all
Find out and be compared one by one.
Retrieval flow refers to enter the flow process of line retrieval in flow process storehouse, and related procedure is referred to be existed with retrieval flow
The flow process of certain contact.One retrieval flow will find out flow process associated with it in flow process storehouse, it is thus necessary to determine that retrieval flow and
Degree of correlation in flow process storehouse between all flow processs.Prior art judges whether correlation is by measuring between two flow processs to two flow processs
Similarity carrying out, for example, give threshold value, similarity is otherwise uncorrelated more than the threshold value then two flow process correlations.
The method of discrimination of existing flow process correlation only considered a dimension mostly, but in fact the degree of correlation between flow process needs
It is analyzed from different dimensions.Also, the method for discrimination of existing flow process correlation is in retrieval flow and flow process storehouse is calculated
After similarity between all flow processs, the similarity of all of flow process conjunctive search flow process in flow process storehouse is ranked up from high to low,
The related procedure that top n flow process is the retrieval flow is generally only provided, and user sometimes for obtain whole related procedures with
The correlation of retrieval flow.
The content of the invention
In view of this, the invention solves the problems that a technical problem be to provide a kind of method of discrimination of operation flow correlation
And device.
According to one embodiment of present invention, there is provided a kind of method of discrimination of operation flow correlation, including：Determine flow process
Sample, and determine the business weight feature for measuring correlation between flow process, wherein, business weight feature includes：Architectural feature
With behavior relation feature；Operation flow in calculation process sample and flow process storehouse is for the linked character of the business weight feature
Value；It is determined that the associated objects value whether similar for identifying the flow process sample and the operation flow；It is special based on the association
Value indicative and the associated objects value generate classification based training relation vector, and according to the classification based training relation vector classification based training is carried out
Obtain disaggregated model；The correlation of new business flow process and the operation flow is differentiated based on the disaggregated model.
Alternatively, the flow process sample and the operation flow are patterned business process model, including：Node, use
In connecting and have directive side by the node；The architectural feature includes：Node is replaced, deleted/is inserted node, replace
While, delete/insertion while；The behavior relation feature includes：Ordinal relation, mutex relation, concurrency relation.
Alternatively, the flow process sample and the operation flow in flow process storehouse are calculated for the association of the business weight feature
Characteristic value includes：It is determined that for replace node or deletion/insertion node linked character value fv1=(| f1 (Process1,
Process2)|)/(|Node(Process 1)∪Node(Process2)|)；Wherein, Process1 is the flow process sample,
Process2 is the operation flow, and f1 (Process 1, Process 2) is replaced for identical in Process1 and Process2
Change the element number of the set of node or deletion/insertion node, Node (Process 1) for Process 1 in all nodes
Set element number, Node (Process 2) for Process 2 in all nodes set in element number, 0≤
fv1≤1。
Alternatively, the flow process sample and the operation flow in flow process storehouse are calculated for the association of the business weight feature
Characteristic value includes：It is determined that for replace while or linked character value fv2=during deletion/insertion (| f2 (Process 1,
Process2)|)/(|Edge(Process1)∪Edge(Process2)|)；Wherein, Process1 is the flow process sample,
Process2 is the operation flow, and f2 (Process 1, Process 2) is replaced for identical in Process1 and Process2
Change sides or delete/insert the element number of the set on side, Edge (Process 1) for Process 1 in all sides set
Element number, Edge (Process 2) for Process 2 in all sides set element number, 0≤fv2≤1.
Alternatively, the flow process sample and the operation flow in flow process storehouse are calculated for the association of the business weight feature
Characteristic value includes：It is determined that for ordinal relation, mutex relation or concurrency relation linked character value fv3=(| f3 (Process
1)∩f3(Process 2)|)/(|f3(Process1)∪f3(Process 2)|)；Wherein, Process1 is the flow process sample
This, Process2 is the operation flow, and f3 (Process 1) is the ordinal relation in Process 1, mutex relation or concurrent
The element number of the set of relation, Edge (Process 2) is the ordinal relation in Process 2, mutex relation or concurrently close
The element number of the set of system, 0≤fv3≤1.
Alternatively, generating classification based training relation vector based on the linked character value and the associated objects value includes：If
Put flow process sample set, calculate each flow process sample in the flow process sample set respectively with operation flow storehouse in each Business Stream
Journey is for the linked character value of business weight feature each described；Determine each flow process sample respectively with the pass of each operation flow
Connection desired value；Generate each flow process sample respectively with the classification based training relation vector of each operation flow, wherein institute
Stating the element of classification based training relation vector includes：The linked character value, the associated objects value.
Alternatively, differentiate that new business flow process includes with the correlation of the operation flow by the disaggregated model：Pass through
Disaggregated model obtains the relating value of each operation flow that the new business flow process is concentrated with the operation flow, based on the pass
Connection value judges whether correlation.
Alternatively, the patterned business process model includes：Petri net model；The associated objects value is 1 or 0,
Wherein, 1 shows to be associated, or 0 shows dereferenced.
According to a further aspect in the invention, there is provided a kind of discriminating gear of operation flow correlation, including：Sample arranges mould
Block, for determining flow process sample；Characteristic determination module, for determining the business weight feature for measuring correlation between flow process,
Wherein, the business weight feature includes：Architectural feature and behavior relation feature；Characteristic value calculating module, it is described for calculating
Operation flow in flow process sample and flow process storehouse is for the linked character value of the business weight feature；Desired value determining module,
For determining for identifying the whether similar associated objects value of the flow process sample and the operation flow；Model training module,
For generating classification based training relation vector based on the linked character value and the associated objects value, closed according to the classification based training
It is that vector carries out classification based training acquisition disaggregated model；Association discrimination module, for differentiating new service flow based on the disaggregated model
The correlation of journey and the operation flow.
Alternatively, the flow process sample and the operation flow are patterned business process model, including：Node, use
In connecting and have directive side by the node；The architectural feature includes：Node is replaced, deleted/is inserted node, replace
While, delete/insertion while；The behavior relation feature includes：Ordinal relation, mutex relation, concurrency relation.
Alternatively, the characteristic value calculating module, is additionally operable to determine for the association for replacing node or deletion/insertion node
Characteristic value fv1=(| f1 (Process1, Process2) |)/(| Node (Process1) ∪ Node (Process 2) |)；Its
In, Process1 is the flow process sample, and Process2 is the operation flow, and f1 (Process 1, Process 2) is
Identical replaces the element number of the set of node or deletion/insertion node, Node in Process1 and Process2
(Process 1) for Process1 in all nodes set element number, Node (Process 2) is Process 2
In all nodes set in element number, 0≤fv1≤1.
Alternatively, the characteristic value calculating module, be additionally operable to determine for replace while or linked character during deletion/insertion
Value fv2=(| f2 (Process 1, Process2) |)/(| Edge (Process1) ∪ Edge (Process 2) |)；Wherein,
Process1 is the flow process sample, and Process2 is the operation flow, and f2 (Process 1, Process 2) is
In Process1 and Process2 identical replace while or set during deletion/insertion element number, Edge (Process 1)
The element number of the set on all sides in for Process 1, Edge (Process 2) for Process 2 in all sides
The element number of set, 0≤fv2≤1.
Alternatively, the characteristic value calculating module, is additionally operable to determine for ordinal relation, mutex relation or concurrency relation
Linked character fv3=(| f3 (Process 1) ∩ f3 (Process 2) |)/(| f3 (Process 1) ∪ f3 (Process 2)
|)；Wherein, Process1 is the flow process sample, and Process2 is the operation flow, and f3 (Process 1) is Process
The element number of the set of ordinal relation, mutex relation or concurrency relation in 1, Edge (Process 2) is in Process 2
Ordinal relation, mutex relation or concurrency relation set element number, 0≤fv3≤1.
Alternatively, the sample setup module, is additionally operable to setting procedure sample set；The characteristic value calculating module, is used for
Calculate each flow process sample in the flow process sample set respectively with operation flow storehouse in each operation flow for each institute
State the linked character value of business weight feature；The model training module, for determine each flow process sample respectively with each industry
The associated objects value of business flow process；Generate each flow process sample respectively with the classification based training relation of each operation flow to
Amount, wherein the element of the classification based training relation vector includes：The linked character value, the associated objects value.
Alternatively, the association discrimination module, is additionally operable to obtain the new business flow process and the industry by disaggregated model
The relating value of each operation flow that business flow process is concentrated, based on the relating value correlation is judged whether.
Alternatively, the patterned business process model includes：Petri net model；The associated objects value is 1 or 0,
Wherein, 1 shows to be associated, or 0 shows dereferenced.
The method of discrimination and device of the operation flow correlation of the present invention, in correlation judgement structure and behavior has been merged
Two dimensions, determine the business weight feature and the linked character value of calculation process of representative structure and behavior dimension, based on association
Characteristic value and associated objects value carry out classification based training and generate disaggregated model, and by disaggregated model new business flow process and Business Stream are differentiated
The correlation of journey, by the similarity retrieval for flow process pretreatment work is carried out, and can improve similarity retrieval and correlation
Property judgement accuracy rate and efficiency.
Description of the drawings
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
The accompanying drawing to be used needed for having technology description is briefly described, it should be apparent that, drawings in the following description are only
Some embodiments of the present invention, for those of ordinary skill in the art, without having to pay creative labor, also
Other accompanying drawings can be obtained according to these accompanying drawings.
Fig. 1 is the schematic flow sheet of one embodiment of the method for discrimination of the operation flow correlation according to the present invention；
Fig. 2A is the flow process sample schematic diagram modeled using Petri net model, and Fig. 2 B are to be modeled using Petri net model
Operation flow schematic diagram；
Fig. 3 A are characterized the schematic diagram of vector, and Fig. 3 B are the schematic diagram of classification based training relation vector；
Fig. 4 is the schematic flow sheet of another embodiment of the method for discrimination of the operation flow correlation according to the present invention；
Fig. 5 is the module diagram of one embodiment of the discriminating gear of the operation flow correlation according to the present invention.
Specific embodiment
The present invention is described more fully with reference to the accompanying drawings, wherein illustrating the exemplary embodiment of the present invention.Under
Face is clearly and completely described with reference to the accompanying drawing in the embodiment of the present invention to the technical scheme in the embodiment of the present invention, shows
So, described embodiment is only a part of embodiment of the invention, rather than the embodiment of whole.Based on the reality in the present invention
Example is applied, the every other embodiment that those of ordinary skill in the art are obtained under the premise of creative work is not made all belongs to
In the scope of protection of the invention.Many descriptions are carried out to technical scheme with reference to each figure and embodiment.
Fig. 1 is the schematic flow sheet of one embodiment of the method for discrimination of the operation flow correlation according to the present invention, such as
Shown in Fig. 1：
Step 101, determines flow process sample, and determines the business weight feature for measuring correlation between flow process, operational authority
Weight feature includes：Architectural feature and behavior relation feature.
Step 102, the operation flow in calculation process sample and flow process storehouse is for the linked character value of business weight feature.
Step 103, it is determined that the associated objects value whether similar for identifying flow process sample and operation flow.
Step 104, based on linked character value and associated objects value classification based training relation vector is generated, and is closed according to classification based training
It is that vector carries out classification based training acquisition disaggregated model.
Step 105, based on disaggregated model the correlation of new business flow process and operation flow is differentiated.
Flow process sample and operation flow can be patterned business process model, including：Node, for node to be connected
And have a directive side.Acquisition flow process sample, operation flow etc. can be modeled using various models.As shown in Fig. 2A, 2B,
Process 1 and Process 2 be with Petri network model procedural model, wherein square nodes represent transition, represent task or
A in event, such as Process 1, B, C, D, E, F node；Circular node represents place, represents state or condition, such as Process
P1 in 1, P2, P3, P4, P5, P6, P7 node；Directed arc is the side connected between place and transition, in Process 1
P1 → A, A → P2.
Degree of correlation between two flow processs of tolerance generally considers following dimension：
(1) node label similarity, i.e., based on corresponding task node between two procedural models and other node identifications it
Between similarity.Essential element is various nodes in procedural model, and the tag name of node can give expression to a certain extent the stream
The correlation function and application of Cheng Jiedian.
(2) structural similarity：Each service logic unit formula in the topological structure major embodiment of flow process flow process passes through
What which kind of logical relation was connected with each other, it determines to a great extent associated traffic data and controlling stream in flow process
Rotation direction.
(3) behavior similarity：The behavior of flow process best embodies in actual motion each movable dependence in flow process.
It is attached the representation of flow process with which kind of topological structure to node and side, is a weight of flow process
Want feature.Figure editing distance is a kind of good mode for weighing structural similarity, mainly from replacement node, replacement side, deletion
Node, six features are deleted in, insertion node, insertion considering.It is of the invention to consider from two dimensions of structure and behavior, it is determined that
Architectural feature include：Replace node, delete/insertion node, replace while, delete/insertion while.Behavior relation feature includes：It is suitable
Order relation, mutex relation, concurrency relation.
Replace node：Two nodes in two flow charts are to replace node, then mean that this has certain to node
Plant corresponding relation.General this corresponding relation is all the similitude between node label, can be with text similarity or semantic phase
Calculate like degree.Similitude between node label is bigger, then this is more likely to become replacement node to node.Delete/insertion section
Point：After replacement node in two flow processs are identified, remaining node is exactly the node deleted or insert.
Replace side：In flow charts, the left and right two ends of a line correspond to respectively two nodes, referred to as left sibling and right section
Point.If two sides in two flow processs are to replace, the left sibling on this two sides is to replace node and right node
It is to replace node.Delete/insertion side：In flow chart when being to delete or insert, then the corresponding left sibling in the side is not replaced
Node is changed, or right node is not to replace node, or left sibling and right node are not to replace node.
Behavior is the critically important dimension of flow chart, and it is with assorted to essentially describe between each task node in a flow process
What the relation order of sample was performed：Sequentially, (mutual exclusion), concurrency relation are selected.Ordinal relation：Two task sections in one flow process
The execution relation of point is order, that is, one of task is first carried out, and another task will be completed in this tasks carrying
After could perform.(mutual exclusion) relation of selection：If two task nodes in a flow process are choice relations, it means that flow process
A node in two task nodes can be selected to perform and do not perform another.In the flow process modeled with Petri network, two
Task node (transition) is choice relation, then they have common forerunner's place, and two branch roads are subsequently divided at once.It is parallel to close
System：If the execution relation between two task nodes in a flow process is neither ordinal relation is nor choice relation, it
Between be concurrency relation.Namely can perform simultaneously between two task nodes.
In one embodiment, two flow processs " Process 1 " and " Process 2 " correspondence replace node or deletion/insert
The linked character value of ingress：
Fv1=(| f1 (Process1, Process2) |)/(| Node (Process 1) ∪ Node (Process2) |)；
(1)
Process1 is flow process sample, and Process2 is operation flow, and f1 (Process 1, Process 2) is
Identical replaces the element number of the set of node or deletion/insertion node, Node in Process1 and Process2
(Process 1) for Process 1 in all nodes set element number, Node (Process 2) is Process 2
In all nodes set in element number, 0≤fv1≤1.
It is determined that for replace while or linked character value during deletion/insertion：
Fv2=(| f2 (Process 1, Process2) |)/(| Edge (Process1) ∪ Edge (Process2) |)；
(2)
F2 (Process 1, Process 2) replaces side or deletion/insertion for identical in Process1 and Process2
The element number of the set on side, Edge (Process 1) for Process 1 in all sides set element number, Edge
(Process 2) for Process 2 in all sides set element number, 0≤fv2≤1.
Each flow process has factum, therefore the behavioral similarity of two flow processs can be with behavior common factor number size
To weigh.It is determined that for the linked character value of ordinal relation, mutex relation or concurrency relation：
Fv3=(| f3 (Process 1) ∩ f3 (Process 2) |)/(| f3 (Process 1) ∪ f3 (Process 2)
|)(3)
F3 (Process 1) is individual for the element of the set of the ordinal relation in Process 1, mutex relation or concurrency relation
Number, Edge (Process 2) for the set of the ordinal relation in Process 2, mutex relation or concurrency relation element number,
0≤fv3≤1.Can see from formula (3), two flow processs depend on their behaviors simultaneously for the characteristic value of same feature
The common factor number and union number of feature.
The replacement node of Process 1 and Process 2 is { A, B, C, D, P1, P2, P3, P4, P5, P6 }, wherein A, B,
C, D } it is the transition replaced, { P1, P2, P3, P4, P5, P6, P7 } is the place replaced.Delete/insertion node for E, P7, F, G,
P8, H }.
Replace side for P1 → A, A → P2, P2 → B, B → P3, P3 → D, D → P6, A → P4, P4 → C, C → P5, P5 →
D}.It is { P1 → E, E → P7, P7 → F, F → P6, P1 → G, G → P8, P8 → H, H → P6 } to delete/insert side.
Ordinal relation has { A → B, A → D, A → C, B → D, C → D, E → F }, the ordinal relation in Process 2 have A →
B, A → D, A → C, B → D, C → D, G → H }, wherein " → " represents the ordinal relation of two nodes.Choice relation has { A#E, A#
F, B#E, B#F, C#E, C#F, D#E, D#F }, { A#G, A#H, B#G, B#H, C#G, C#H, D#G, D#H } is the selection of Process 2
Relation, wherein " # " represent that between two task nodes be choice relation.Concurrency relation is all { B==C }, wherein "==" refer to
Be concurrency relation.
Calculated according to formula (1), the replacement nodal value of Process 1 and Process 2 be fv1 (replacement node)=
10/16, deletion/insertion nodal value is fv1 (deleting/insertion node)=6/16.
Calculated according to formula (2), the replacement boundary values of Process 1 and Process 2 is fv2 (replacement side)=10/
18, deletion/insertion nodal value is fv2 (deleting/insertion side)=8/18.
Calculated according to formula (3), ordinal relation value is fv3 between the task node of Process 1 and Process 2
(order)=5/7, choice relation value is fv3 (selection)=0, and concurrency relation value is fv3 (parallel)=1.
The more big then corresponding feature of characteristic value is more similar or more dissimilar.As shown in Fig. 3 A, 3B, feature based value, definition
Two data structures：Characteristic vector and relation vector.Characteristic vector FV=of Process 1 and Process 2 (replaces node
Value, deletes/insertion nodal value, replaces boundary values, deletes/insertion boundary values, ordinal relation value, choice relation value, concurrency relation value),
By 7 eigenvalue clusters into dependency relation one characteristic vector of correspondence between two flow processs.Relation vector is in characteristic vector
On the basis of be worth to plus target, relation vector RV=(characteristic value, desired value).In order to weigh the classification of sorting algorithm
Accuracy rate, manually to two flow processs, whether correlation provides a desired value when data are prepared, and the value is 1 if correlation,
Otherwise it is 0.
The relation vector of Process 1 and Process 2=(characteristic vector, desired value).Artificially judge Process 1
It is related to Process 2, so desired value is 1.Thus, Process 1 and the corresponding characteristic vectors of Process 2 are FV=
(0.625,0.375,0.556,0.444,0.713,0,1).
Setting procedure sample set, each the flow process sample in calculation process sample set respectively with operation flow storehouse in each
Operation flow is for the linked character value of each business weight feature.Determine each flow process sample respectively with each operation flow
Associated objects value, generate each flow process sample respectively with the classification based training relation vector of each operation flow, wherein classification based training
The element of relation vector includes：Linked character value, associated objects value.
Classification based training can be carried out using many algorithms model, obtain sub-model, for example, BP neural network algorithm, vector
Support machine (SVM) algorithm etc..The pass of each operation flow that new business flow process is concentrated with operation flow is obtained by disaggregated model
Connection value, based on relating value correlation is judged whether.
In one embodiment, in order to adapt to new demand, the Process Designer of certain telecommunications enterprise is continuous to cellphone subscriber
About flow process is redesigned.Before new flow process is added to into flow process storehouse, in order to avoid redundancy, need first in flow process storehouse
Related procedure all find out and be compared one by one.In flow process storehouse, new flow process will be carried out one by one with existing operation flow
Comparison, if certain related procedure and new technological process belong to the flow process of same type, the operation flow will be removed from flow process storehouse,
In case redundancy.
Fig. 4 is the schematic flow sheet of another embodiment of the method for discrimination of the operation flow correlation according to the present invention,
As shown in Figure 4
Step 401, gives m retrieval flow Q1..., QmN prioritizing C is included with one1..., CnFlow process
Storehouse.
Previously given 10 retrieval flows are used as flow process sample.For example, Q1：Reservation flow process, Q2：Bill querying flow, Q3：
Integration querying flow, Q4：Accumulated point exchanging gift flow process, Q5：Bill payment flow process, Q6：Telephone recharge flow process, Q7：Flow orders stream
Journey, Q8：Cellphone subscriber's renewed treaty flow process, Q9：Handle flow process, Q in broadband10：Handle major-minor card flow process.One is provided with comprising 100
Operation flow C1..., C100Telecommunications flow process storehouse, this 100 operation flows be prioritizing
Step 402：Read in 1 retrieval flow Qi(1≤i≤10)。
Step 403：Take out operation flow C in flow process storehousej(1≤j≤100)。
Step 404, calculates QiAnd Cj7 characteristic values, and constitute a characteristic vector FV (Qi, Cj)=(replaces node
Value, deletes/insertion nodal value, replaces boundary values, deletes/insertion boundary values, ordinal relation value, choice relation value, concurrency relation value).
For example, telephone expenses querying flow and flow inquiry all by sending message search or can be inquired by telephone or done business on the net
The Room is inquired about.Telephone expenses querying flow：Dial XXXX/ (logging in online business hall → click telephone expenses inquiry)/transmission X to XXXX → return
Telephone expenses details；Flow querying flow：Dial XXXX/ (logging in online business hall → click traffic inquiry)/transmission Y to XXXX → return
Capacity of returns details.
The node of two flow process replacements is " logging in online business hall ", " calling XXXX ", and deletion of node is for " click is talked about
Take inquiry ", " send X to XXXX ", " returning telephone expenses details ", insertion node is " click traffic inquiry ", " transmission Y to XXXX ",
" return flow details ".It is 0 that side number is replaced all to replace node, thus due to two nodes on no side, and all sides are all to delete
Except/insertion side.According to these information, telephone expenses querying flow and flow can be obtained according to formula (1), formula (2) and formula (3)
The characteristic vector of querying flow.
Step 405, artificially to QiWith CjCarry out the judgement of correlation and provide desired value, 1 represents related, and 0 represents not phase
Close.For example, telephone expenses querying flow is related to flow inquiry, because being all querying flow, is by the two Process Markups then
Correlation, is labeled as 1.A relation vector RV (Q is constituted on the basis of FV plus a Target valuei, Cj)。
Step 406, by RV (Qi, Cj) it is written to file.
Step 407, judges whether also have the operation flow not calculated in flow process storehouse, if it is, into step 403,
If it is not, then into step 408.
Step 408, judges whether 10 retrieval flows all calculate and finishes.If it is, execution step 403, such as otherwise holds
Row step 408.
Step 409, by the file comprising 1000 records BP neural network (BPNN) or vectorial support machine (SVM) are input to
It is trained in algorithm, obtains disaggregated model.
Step 401-409 is training process under line.
Step 410, gives 1 new retrieval flow Q.For example, cellphone subscriber's renewed treaty flow process, is input to the classification for training
In model.
Step 411, takes out the prioritizing C in flow process storehousej(1≤j≤100)。
Step 412, by disaggregated model to Q and CjWhether correlation is judged, i.e. output 0 or 1,0 represents Q and CjCorrelation,
1 represents Q and CjIt is uncorrelated.
For example, the flow process such as broadband renewed treaty flow process in flow process storehouse is the related procedure of Q, is labeled as 1.Other in flow process storehouse
Flow process, such as troublshooting flow process, length of surfing the Net speed-raising flow process, shutdown guarantor's flow process flow process are the uncorrelated flow process of Q, are labeled as 0.
Step 413, judges in flow process storehouse whether to also have uncalculated operation flow, if it is, into step 410, such as
It is really no, then terminate.
Step 410-413 is on-line checking process.
The method of discrimination of the operation flow correlation provided in above-described embodiment, has merged two dimensions of structure and behavior,
The business weight feature and the linked character value of calculation process of representative structure and behavior dimension are determined, based on linked character value and pass
Connection desired value carries out classification based training and generates disaggregated model, differentiates that new business flow process is related to operation flow by disaggregated model
Property, pretreatment work is carried out by the similarity retrieval for flow process, the accuracy rate and efficiency of similarity retrieval can be improved.
In one embodiment, the present invention provides a kind of discriminating gear 50 of operation flow correlation, including：Sample is arranged
Module 51, characteristic determination module 52, characteristic value calculating module 53, desired value determining module 54, model training module 55 with associate
Discrimination module 56.Sample setup module 51 determines flow process sample；Characteristic determination module 52 is determined for measuring correlation between flow process
Business weight feature, business weight feature includes：Architectural feature and behavior relation feature.
Operation flow in the calculation process sample of characteristic value calculating module 53 and flow process storehouse is for the pass of business weight feature
Connection characteristic value.Desired value determining module 54 determines the associated objects value whether similar for identifying flow process sample and operation flow.
Model training module 55 generates classification based training relation vector based on linked character value and associated objects value, according to classification based training relation
Vector carries out classification based training and obtains disaggregated model.Association discrimination module 56 differentiates new business flow process and Business Stream based on disaggregated model
The correlation of journey.
Characteristic value calculating module 53 determine for replace node or deletion/insertion node linked character value fv1=(| f1
(Process1, Process2) |)/(| Node (Process 1) ∪ Node (Process2) |)；Wherein, Process1 is flow process
Sample, Process2 is operation flow, and f1 (Process1, Process 2) is replaced for identical in Process1 and Process2
Change the element number of the set of node or deletion/insertion node, Node (Process 1) for Process 1 in all nodes
Set element number, Node (Process 2) for Process 2 in all nodes set in element number, 0≤
fv1≤1。
Characteristic value calculating module 53 determine for replace while or linked character value fv2=during deletion/insertion (| f2
(Process 1, Process2) |)/(| Edge (Process1) ∪ Edge (Process 2) |)；Wherein, Process1 is stream
Journey sample, Process2 is operation flow, and f2 (Process 1, Process 2) is identical in Process1 and Process2
Replace while or set during deletion/insertion element number, Edge (Process 1) for Process 1 in all sides collection
The element number of conjunction, Edge (Process 2) for Process 2 in all sides set element number, 0≤fv2≤1.
Characteristic value calculating module 53 determine for ordinal relation, mutex relation or concurrency relation linked character fv3=(|
f3(Process 1)∩f3(Process 2)|)/(|f3(Process 1)∪f3(Process2)|)；Wherein, Process1
For flow process sample, Process2 is operation flow, f3 (Process 1) is the ordinal relation in Process 1, mutex relation or
The element number of the set of concurrency relation, Edge (Process 2) is the ordinal relation in Process 2, mutex relation or and
The element number of the set of the relation of sending out, 0≤fv3≤1.
The setting procedure sample set of sample setup module 51.Each stream in the calculation process sample set of characteristic value calculating module 53
Journey sample respectively with operation flow storehouse in each operation flow for the linked character value of each business weight feature.Model is instructed
Practice module 54 determine each flow process sample respectively with the associated objects value of each operation flow, generate each flow process sample respectively with
The classification based training relation vector of each operation flow, wherein, the element of classification based training relation vector includes：Linked character value, pass
Connection desired value.Association discrimination module 55 obtains each operation flow that new business flow process is concentrated with operation flow by disaggregated model
Relating value, correlation is judged whether based on relating value.
The method of discrimination and device of the operation flow correlation provided in above-described embodiment, has merged structure and behavior two
Dimension, determines the business weight feature and the linked character value of calculation process of representative structure and behavior dimension, based on linked character
Value and associated objects value carry out classification based training and generate disaggregated model, differentiate new business flow process with operation flow by disaggregated model
Correlation, can be divided into related to the new business flow process and uncorrelated big class of flow process two, by right by the operation flow in flow process storehouse
Pretreatment work is carried out in the similarity retrieval of flow process, the accuracy rate and effect of similarity retrieval and correlation judgement can be improved
Rate.
The method of the present invention and system may be achieved in many ways.For example, can by software, hardware, firmware or
Software, hardware, any combinations of firmware are realizing the method for the present invention and system.Only it is for said sequence the step of method
In order to illustrate, order described in detail above is not limited to the step of the method for the present invention, especially say unless otherwise
It is bright.Additionally, in certain embodiments, also the present invention can be embodied as recording program in the recording medium, these programs include
For realizing the machine readable instructions of the method according to the invention.Thus, the present invention also covers storage for performing according to this
The recording medium of the program of bright method.
Description of the invention is given for the sake of example and description, and is not exhaustively or by the present invention
It is limited to disclosed form.Many modifications and variations are for the ordinary skill in the art obvious.Select and retouch
It is to more preferably illustrate the principle and practical application of the present invention, and one of ordinary skill in the art is managed to state embodiment
The present invention is solved so as to design the various embodiments with various modifications for being suitable to special-purpose.
Claims (16)
1. a kind of method of discrimination of operation flow correlation, it is characterised in that include：
Determine flow process sample, and determine the business weight feature for measuring correlation between flow process, wherein, the business weight is special
Levy including：Architectural feature and behavior relation feature；
The flow process sample and the operation flow in flow process storehouse are calculated for the linked character value of the business weight feature；
It is determined that the associated objects value whether similar for identifying the flow process sample and the operation flow；
Classification based training relation vector is generated based on the linked character value and the associated objects value, is closed according to the classification based training
It is that vector carries out classification based training acquisition disaggregated model；
The correlation of new business flow process and the operation flow is differentiated based on the disaggregated model.
2. the method for claim 1, it is characterised in that
The flow process sample and the operation flow are patterned business process model, including：Node, for by the node
Connect and have directive side；
The architectural feature includes：Replace node, delete/insertion node, replace while, delete/insertion while；
The behavior relation feature includes：Ordinal relation, mutex relation, concurrency relation.
3. method as claimed in claim 2, it is characterised in that calculate the flow process sample and the operation flow pair in flow process storehouse
Include in the linked character value of the business weight feature：
It is determined that for the linked character value for replacing node or deletion/insertion node
Fv1=(| f1 (Process1, Process2) |)/(| Node (Process1) ∪ Node (Process2) |)；
Wherein, Process1 be the flow process sample, Process2 be the operation flow, f1 (Process1, Process2)
The element number of the set of node or deletion/insertion node, Node are replaced for identical in Process1 and Process2
(Process1) for all nodes in Process1 set element number, Node (Process2) is in Process2
Element number in the set of all nodes, 0≤fv1≤1.
4. method as claimed in claim 2, it is characterised in that calculate the flow process sample and the operation flow pair in flow process storehouse
Include in the linked character value of the business weight feature：
It is determined that for replace while or linked character value during deletion/insertion
Fv2=(| f2 (Process1, Process2) |)/(| Edge (Process1) ∪ Edge (Process2) |)；
Wherein, Process1 be the flow process sample, Process2 be the operation flow, f2 (Process1, Process2)
For identical in Process1 and Process2 replace while or set during deletion/insertion element number, Edge
(Process1) for all sides in Process1 set element number, Edge (Process2) is the institute in Process2
There are the element number of the set on side, 0≤fv2≤1.
5. method as claimed in claim 2, it is characterised in that calculate the flow process sample and the operation flow pair in flow process storehouse
Include in the linked character value of the business weight feature：
It is determined that for the linked character value of ordinal relation, mutex relation or concurrency relation
Fv3=(| f3 (Process1) ∩ f3 (Process2) |)/(| f3 (Process1) ∪ f3 (Process2) |)；
Wherein, Process1 is the flow process sample, and Process2 is the operation flow, and f3 (Process1) is Process1
In ordinal relation, mutex relation or concurrency relation set element number, Edge (Process2) be Process2 in
The element number of the set of ordinal relation, mutex relation or concurrency relation, 0≤fv3≤1.
6. method as claimed in claim 2, it is characterised in that generated based on the linked character value and the associated objects value
Classification based training relation vector includes：
Setting procedure sample set, calculate each flow process sample in the flow process sample set respectively with the operation flow storehouse in
Each operation flow is for the linked character value of business weight feature each described；
Determine each flow process sample respectively with the associated objects value of each operation flow；
Generate each flow process sample respectively with the classification based training relation vector of each operation flow, wherein the classification
The element of training relation vector includes：The linked character value, the associated objects value.
7. method as claimed in claim 6, it is characterised in that new business flow process and the industry are differentiated by the disaggregated model
The correlation of business flow process includes：
The relating value of each operation flow that the new business flow process is concentrated with the operation flow is obtained by disaggregated model, and
Correlation is judged whether based on the relating value.
8. method as claimed in claim 2, it is characterised in that
The patterned business process model includes：Petri net model；
The associated objects value is 1 or 0, wherein, 1 shows to be associated, or 0 shows dereferenced.
9. a kind of discriminating gear of operation flow correlation, it is characterised in that include：
Sample setup module, for determining flow process sample；
Characteristic determination module, for determining the business weight feature for measuring correlation between flow process, wherein, the business weight
Feature includes：Architectural feature and behavior relation feature；
Characteristic value calculating module, for calculating the flow process sample and operation flow in flow process storehouse is for the business weight is special
The linked character value levied；
Desired value determining module, for determining for identifying the whether similar association mesh of the flow process sample and the operation flow
Scale value；
Model training module, for generating classification based training relation vector based on the linked character value and the associated objects value,
Classification based training is carried out according to the classification based training relation vector and obtains disaggregated model；
Association discrimination module, for differentiating the correlation of new business flow process and the operation flow based on the disaggregated model.
10. device as claimed in claim 9, it is characterised in that
The flow process sample and the operation flow are patterned business process model, including：Node, for by the node
Connect and have directive side；
The architectural feature includes：Replace node, delete/insertion node, replace while, delete/insertion while；The behavior relation is special
Levy including：Ordinal relation, mutex relation, concurrency relation.
11. devices as claimed in claim 10, it is characterised in that
The characteristic value calculating module, is additionally operable to determine for linked character value fv1=for replacing node or deletion/insertion node
(| f1 (Process1, Process2) |)/(| Node (Process1) ∪ Node (Process2) |)；
Wherein, Process1 be the flow process sample, Process2 be the operation flow, f1 (Process1, Process2)
The element number of the set of node or deletion/insertion node, Node are replaced for identical in Process1 and Process2
(Process1) for all nodes in Process1 set element number, Node (Process2) is in Process2
Element number in the set of all nodes, 0≤fv1≤1.
12. devices as claimed in claim 10, it is characterised in that
The characteristic value calculating module, be additionally operable to determine for replace while or linked character value fv2=during deletion/insertion (| f2
(Process1, Process2) |)/(| Edge (Process1) ∪ Edge (Process2) |)；
Wherein, Process1 be the flow process sample, Process2 be the operation flow, f2 (Process1, Process2)
For identical in Process1 and Process2 replace while or set during deletion/insertion element number, Edge
(Process1) for all sides in Process1 set element number, Edge (Process2) is the institute in Process2
There are the element number of the set on side, 0≤fv2≤1.
13. devices as claimed in claim 10, it is characterised in that
The characteristic value calculating module, is additionally operable to determine for the linked character fv3 of ordinal relation, mutex relation or concurrency relation
=(| f3 (Process1) ∩ f3 (Process2) |)/(| f3 (Process1) ∪ f3 (Process2) |)；
Wherein, Process1 is the flow process sample, and Process2 is the operation flow, and f3 (Process1) is Process1
In ordinal relation, mutex relation or concurrency relation set element number, Edge (Process2) be Process2 in
The element number of the set of ordinal relation, mutex relation or concurrency relation, 0≤fv3≤1.
14. devices as claimed in claim 10, it is characterised in that
The sample setup module, is additionally operable to setting procedure sample set,
The characteristic value calculating module, be additionally operable to calculate each flow process sample in the flow process sample set respectively with operation flow
Each operation flow in storehouse is for the linked character value of business weight feature each described；
The model training module, for determine each flow process sample respectively with the associated objects value of each operation flow；Generate
Each flow process sample respectively with the classification based training relation vector of each operation flow, wherein the classification based training relation
The element of vector includes：The linked character value, the associated objects value.
15. devices as claimed in claim 14, it is characterised in that
The association discrimination module, is additionally operable to obtain what the new business flow process and the operation flow were concentrated by disaggregated model
The relating value of each operation flow, based on the relating value correlation is judged whether.
16. devices as claimed in claim 10, it is characterised in that
The patterned business process model includes：Petri net model；
The associated objects value is 1 or 0, wherein, 1 shows to be associated, or 0 shows dereferenced.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611231847.6A CN106651317A (en) | 2016-12-28 | 2016-12-28 | Method and device for judging business process correlation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611231847.6A CN106651317A (en) | 2016-12-28 | 2016-12-28 | Method and device for judging business process correlation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106651317A true CN106651317A (en) | 2017-05-10 |
Family
ID=58832927
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611231847.6A Pending CN106651317A (en) | 2016-12-28 | 2016-12-28 | Method and device for judging business process correlation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106651317A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108229909A (en) * | 2017-12-14 | 2018-06-29 | 四川虹慧云商科技有限公司 | A kind of resident's transaction methods |
| CN109598281A (en) * | 2018-10-11 | 2019-04-09 | 阿里巴巴集团控股有限公司 | A kind of business risk preventing control method, device and equipment |
| CN111461566A (en) * | 2020-04-10 | 2020-07-28 | 武汉大学 | Cross-boundary service flow fusion method and system based on message flow division and combination |
| CN113240389A (en) * | 2021-05-12 | 2021-08-10 | 网宿科技股份有限公司 | Workflow creating method, electronic device and storage medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050278139A1 (en) * | 2004-05-28 | 2005-12-15 | Glaenzer Helmut K | Automatic match tuning |
| US20110178965A1 (en) * | 2006-11-09 | 2011-07-21 | Pucher Max J | Method for training a system to specifically react on a specific input |
| CN102542406A (en) * | 2011-12-15 | 2012-07-04 | 北京航空航天大学 | Trans-granularity business process inquiring method and device |
| CN103927613A (en) * | 2014-03-13 | 2014-07-16 | 浙江大学 | Process retrieval system and method |
| CN104134112A (en) * | 2014-07-14 | 2014-11-05 | 安徽理工大学 | Business process model consistency measurement method under semantic constraints |
| CN105069574A (en) * | 2015-08-17 | 2015-11-18 | 安徽理工大学 | New method for analyzing business flow behavior similarity |
| CN105184486A (en) * | 2015-09-08 | 2015-12-23 | 国网智能电网研究院 | Power grid business classification method based on directed acyclic graphs support vector machine |
| CN105184316A (en) * | 2015-08-28 | 2015-12-23 | 国网智能电网研究院 | Support vector machine power grid business classification method based on feature weight learning |
| CN105574538A (en) * | 2015-12-10 | 2016-05-11 | 小米科技有限责任公司 | Classification model training method and apparatus |
-
2016
- 2016-12-28 CN CN201611231847.6A patent/CN106651317A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050278139A1 (en) * | 2004-05-28 | 2005-12-15 | Glaenzer Helmut K | Automatic match tuning |
| US20110178965A1 (en) * | 2006-11-09 | 2011-07-21 | Pucher Max J | Method for training a system to specifically react on a specific input |
| CN102542406A (en) * | 2011-12-15 | 2012-07-04 | 北京航空航天大学 | Trans-granularity business process inquiring method and device |
| CN103927613A (en) * | 2014-03-13 | 2014-07-16 | 浙江大学 | Process retrieval system and method |
| CN104134112A (en) * | 2014-07-14 | 2014-11-05 | 安徽理工大学 | Business process model consistency measurement method under semantic constraints |
| CN105069574A (en) * | 2015-08-17 | 2015-11-18 | 安徽理工大学 | New method for analyzing business flow behavior similarity |
| CN105184316A (en) * | 2015-08-28 | 2015-12-23 | 国网智能电网研究院 | Support vector machine power grid business classification method based on feature weight learning |
| CN105184486A (en) * | 2015-09-08 | 2015-12-23 | 国网智能电网研究院 | Power grid business classification method based on directed acyclic graphs support vector machine |
| CN105574538A (en) * | 2015-12-10 | 2016-05-11 | 小米科技有限责任公司 | Classification model training method and apparatus |
Non-Patent Citations (3)
| Title |
|---|
| MICHAEL BECKER: ""A Comparative Survey of Business Process Similarity Measures"", 《COMPUTER IN INDUSTRY》 * |
| 刘海滨: ""以业务单据为中心的业务流程模型聚类及相似性查询方法"", 《计算机集成制造系统》 * |
| 李丹: ""基于流聚类的网络业务识别关键技术研究"", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108229909A (en) * | 2017-12-14 | 2018-06-29 | 四川虹慧云商科技有限公司 | A kind of resident's transaction methods |
| CN109598281A (en) * | 2018-10-11 | 2019-04-09 | 阿里巴巴集团控股有限公司 | A kind of business risk preventing control method, device and equipment |
| CN109598281B (en) * | 2018-10-11 | 2023-05-26 | 创新先进技术有限公司 | Service risk prevention and control method, device and equipment |
| CN111461566A (en) * | 2020-04-10 | 2020-07-28 | 武汉大学 | Cross-boundary service flow fusion method and system based on message flow division and combination |
| CN111461566B (en) * | 2020-04-10 | 2023-02-03 | 武汉大学 | Cross-border service flow fusion method and system based on message flow division and combination |
| CN113240389A (en) * | 2021-05-12 | 2021-08-10 | 网宿科技股份有限公司 | Workflow creating method, electronic device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106651317A (en) | Method and device for judging business process correlation | |
| CN110298574B (en) | Convolutional neural network-based electricity user payment risk rating method | |
| CN107358340A (en) | A kind of project management system based on list of engineering | |
| US20070162482A1 (en) | Method and system of using artifacts to identify elements of a component business model | |
| CN106022708A (en) | Method for predicting employee resignation | |
| CN106095942A (en) | Strong variable extracting method and device | |
| CN111079025A (en) | Government affair recommendation method based on big data analysis and system comprising method | |
| KR102111430B1 (en) | PLM Apparatus and Managing Method Thereof | |
| CN110109902A (en) | A kind of electric business platform recommender system based on integrated learning approach | |
| Zhou et al. | A comprehensive process similarity measure based on models and logs | |
| CN116503158A (en) | Enterprise bankruptcy risk early warning method, system and device based on data driving | |
| CN110188255A (en) | Power consumer Behavior mining method and system based on the shared fusion of business datum | |
| CN112085837B (en) | Three-dimensional model classification method based on geometric shape and LSTM neural network | |
| Zhang et al. | Logistics service supply chain order allocation mixed K-Means and Qos matching | |
| Buss et al. | Volume and capacity interaction in facility design | |
| CN104335161B (en) | Efficient evaluation of network robustness with a graph | |
| Karp | Models and methods for omni-channel fulfillment | |
| CN107093018A (en) | Communication engineering project information method for visualizing and device based on health model | |
| Mon et al. | Evaluation of technological development for the definition of Industries 4.0 | |
| Thotappa et al. | Data mining aided proficient approach for optimal inventory control in supply chain management | |
| CN113034033B (en) | Method for determining variety of newly-researched equipment spare parts | |
| CN110502219B (en) | Enterprise-level IT architecture asset management system and method | |
| CN114662009A (en) | Industrial internet factory collaborative recommendation algorithm based on graph convolution | |
| Salah et al. | Pipeline alignment optimization: Automated GIS-based approach | |
| JP2002215858A (en) | System for creating supply chain model |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170510 |
|
| RJ01 | Rejection of invention patent application after publication |