CN104462288B - A kind of similarity of paths analysis method and system - Google Patents
A kind of similarity of paths analysis method and system Download PDFInfo
- Publication number
- CN104462288B CN104462288B CN201410705903.XA CN201410705903A CN104462288B CN 104462288 B CN104462288 B CN 104462288B CN 201410705903 A CN201410705903 A CN 201410705903A CN 104462288 B CN104462288 B CN 104462288B
- Authority
- CN
- China
- Prior art keywords
- depth value
- node
- subset
- data
- yield
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3409—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment for performance assessment
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3447—Performance evaluation by modeling
Abstract
The embodiment of the invention discloses a kind of similarity of paths analysis method and system.Present invention method includes：The first tree data of reference data and the second tree data of problem data are obtained respectively, and obtains the depth value of each node, and first set and second set are determined according to the depth value of first tree data and each node of second tree data respectively；The similarity weights of the reference data and described problem data are calculated according to the depth value.The present embodiment can by with each node of the first tree data of reference data and the second tree data of problem data by set by way of be indicated, and the similarity weights of the reference data and described problem data are calculated according to each depth value, corresponding relation between reference data and the described problem data obtained from entering, and then utility problem intelligent analysis module can automate performance positioning work, the time of performance issue analysis is greatly improved, the dependence to manpower is reduced.
Description
Technical field
The present invention relates to field of information processing, more particularly to a kind of similarity of paths analysis method and system.
Background technology
In storage service, the module or function passed through by data are different, therefore there are a variety of different I/O paths.
Typically such as in the case where cache is hit, data are directly returned from cache modules；Cache is not hit by data and may proceed to flow direction
Lower module, this is the different I/O paths of two classes.Due to the continuity of system development, between storage system different editions not
Often there is certain corresponding relation with I/O path, such as under same operative scenario, baseline version and later release (are also referred to as problem
Version) I/O path may cause difference due to adding a certain function, empirically think to exist between this two classes I/O path pair
It should be related to, baseline I/O path sports problem I/O path, also referred to as structural mutation.
Under fixed business scenario (such as storage system), the processing path of request (path as read and write IO) be it is fixed,
The change that performance change can trace back to request processing path (causes I/O path elongated, property as read the decline of IO cache hit probabilities
Can accordingly it decline).Change by analysis request path is the effective means of analytical performance change.Path change mainly has two
Kind, 1, time delay mutation：Benchmark I/O path is identical with problem I/O path structure, but the time delay between respective modules has significant difference；
2, structural mutation：Benchmark I/O path is different with problem I/O path structure；For 2, in reference data and problem data, it is necessary to
Maximum possible corresponding relation is calculated in the possibility accidental data of multi-to-multi.For the searching of corresponding relation, currently without generally recognizing
Can method exist.
The current widely used I/O path trace tool of industry has Google dapper, X-Trace etc..Dapper is one
Distributed tracking system is planted, can be with the request call path between tracking server cluster.By taking Dapper as an example, brief description should
Class trace tool principle：
Dapper can track the information of each node in the request paths traversed of user, such as timestamp.Dapper
It is whereby a complete loops by path concatenation using application program or middleware to every request one global flag of record.
Between at the beginning of the track record service of Dapper every and the end time, the ID and father ID of each module, the module without father ID
For root modules, all public tracking ID of tracking are connected in an I/O Request path by tracking ID by these blocks.X-
Trace is as internet host trace tool, and its principle is similar with dapper.Also same principle is used in unified storage system
Trace tool.
These trace tools are all to track sometime point or the I/O path of period, the stable IO roads of such as baseline version
The problem of footpath, problem version I/O path.It can not be solved for the mutation relations problems of the I/O path of different time sections, and due to
The different influences to system between I/O path different levels are different, the larger differences between upper strata IO different reflection versions
It is different, and the IO of lower floor differences may be fine difference between version causes, prior art ignores such difference, to IO not
With being unifiedly calculated, similarity identification reliability is not high, easily causes the erroneous judgement of structural mutation corresponding relation, causes analysis result
Error is larger.
The content of the invention
The embodiments of the invention provide a kind of similarity of paths analysis method and system.
First aspect of the embodiment of the present invention provides a kind of similarity of paths analysis method, including：
The first tree data of reference data and the second tree data of problem data are obtained respectively；
According to the membership between identical traversal rule and each node to first tree data and described
Each node of two tree datas is traveled through to obtain the depth value of each node, and the depth value of each node is used to represent each section
The level that point is located in first tree data or the second tree data；
The first collection is determined according to the depth value of first tree data and each node of second tree data respectively
Close and second set, wherein, the first set includes multiple first subsets corresponding with the depth value respectively, and each described
The depth value of the node of first tree data in first subset is identical, the second set include it is multiple respectively with it is described
The depth value of the node of second tree data in the corresponding yield in the second subset of depth value, and each yield in the second subset is identical；
According to each first subset and depth value corresponding with each yield in the second subset calculate the reference data and
The similarity weights of described problem data.
With reference to the embodiment of the present invention in a first aspect, in the first implementation of the first aspect of the embodiment of the present invention,
The basis is with each first subset and depth value corresponding with each yield in the second subset calculates the base value
Before the similarity weights with described problem data, methods described also includes：
It is determined that the maximum of each depth value corresponding with first subset and determination are corresponding with the yield in the second subset
The maximum of each depth value；
It is determined that the maximum and each depth corresponding with the yield in the second subset of each depth value corresponding with first subset
Smaller value in the maximum of value is the first reference depth value；
It is determined that the maximum and each depth corresponding with the yield in the second subset of each depth value corresponding with first subset
Higher value in the maximum of value is the second reference depth value；
It is determined that initialization similarity weights W is equal to 0；
It is c to determine destination node number, wherein, the c is equal in each first subset and each yield in the second subset
The maximum of nodes.
With reference to the first implementation of the first aspect of the embodiment of the present invention, the of the first aspect of the embodiment of the present invention
In two kinds of implementations,
The basis is with each first subset and depth value corresponding with each yield in the second subset calculates the base value
Include according to the similarity weights with described problem data：
According to each first subset and depth value corresponding with each yield in the second subset order from low to high successively
Determine whether current depth value is less than or equal to first reference depth value；
If so, then determining the first subset corresponding with the current depth value and described according to string editing distance algorithm
SED values between yield in the second subset；
If the SED values are equal to 0, it is determined that the similarity weights W=W+a^(b-1)* 2c, wherein, a span is
It is the current depth value more than 0 and less than 1, b, and b is more than or equal to 1 and less than or equal to first reference depth value；
If the SED values are not equal to 0, it is determined that the similarity weights W=W+a^(b-1)*2c/SED。
With reference to the third implementation of the first aspect of the embodiment of the present invention, the of the first aspect of the embodiment of the present invention
In three kinds of implementations,
The basis is with each first subset and depth value corresponding with each yield in the second subset calculates the base value
Also include according to the similarity weights with described problem data：
If determining that current depth value is more than first reference depth successively according to the order of the depth value from low to high
During value, it is determined that whether the current depth value is less than or equal to second reference depth value；
If, it is determined that the similarity weights W=W-a^(b-1), wherein, b is more than first reference depth value and small
In or equal to second reference depth value.
Second aspect of the embodiment of the present invention provides a kind of system, including：
First acquisition unit, for obtaining the first tree data of reference data and the second tree-like number of problem data respectively
According to；
Second acquisition unit, for according to the membership between identical traversal rule and each node to described first
Each node of tree data and second tree data is traveled through to obtain the depth value of each node, the depth of each node
Angle value is used to represent the level that each node is located in first tree data or the second tree data；
First determining unit, for respectively according to each node of first tree data and second tree data
Depth value determines first set and second set, wherein, the first set includes multiple corresponding with the depth value respectively
The depth value of the node of first tree data in first subset, and each first subset is identical, the second set
Including multiple yield in the second subset corresponding with the depth value respectively, and second tree data in each yield in the second subset
The depth value of node is identical；
Computing unit, for according to each first subset and depth value corresponding with each yield in the second subset calculates institute
State the similarity weights of reference data and described problem data.
With reference to second aspect of the embodiment of the present invention, in the first implementation of second aspect of the embodiment of the present invention,
The system also includes：
Second determining unit, maximum and determination and institute for determining each depth value corresponding with first subset
State the maximum of the corresponding each depth value of yield in the second subset；
3rd determining unit, for determining the maximum of corresponding with first subset each depth value and with described second
Smaller value in the maximum of the corresponding each depth value of subset is the first reference depth value；
4th determining unit, for determining the maximum of corresponding with first subset each depth value and with described second
Higher value in the maximum of the corresponding each depth value of subset is the second reference depth value；
5th determining unit, for determining that initialization similarity weights W is equal to 0；
6th determining unit, for determining that destination node number is c, wherein, the c is equal to each first subset and each
The maximum of nodes in the yield in the second subset.
With reference to the first implementation of second aspect of the embodiment of the present invention, second of second aspect of the embodiment of the present invention
In implementation,
The computing unit includes：
First determining module, for according to each first subset and depth value corresponding with each yield in the second subset by
Low to high order determines whether current depth value is less than or equal to first reference depth value successively；
First computing module, if being less than or equal to first reference depth value for current depth value, according to character
String editing distance algorithm determines the SED values between the first subset corresponding with the current depth value and the yield in the second subset；
Second computing module, if being equal to 0 for the SED values, it is determined that the similarity weights W=W+a^(b-1)* 2c,
Wherein, a span is more than 0 and is the current depth value less than 1, b, and b is more than or equal to 1 and less than or equal to institute
State the first reference depth value；
3rd computing module, if being not equal to 0 for the SED values, it is determined that the similarity weights W=W+a^(b-1)*
2c/SED。
With reference to second of implementation of second aspect of the embodiment of the present invention, second aspect of the embodiment of the present invention the third
In implementation,
The computing unit also includes：
Second determining module, if for determining that current depth value is more than successively according to the order of the depth value from low to high
During first reference depth value, it is determined that whether the current depth value is less than or equal to second reference depth value；
3rd determining module, for if it is determined that the current depth value is less than or equal to second reference depth value, then
Determine the similarity weights W=W-a^(b-1), wherein, b is more than first reference depth value and less than or equal to described second
Reference depth value.
The embodiment of the invention discloses a kind of similarity of paths analysis method and system.Present invention method bag
Include：The first tree data of reference data and the second tree data of problem data are obtained respectively, and obtain the depth of each node
Value, determines first set and the according to the depth value of first tree data and each node of second tree data respectively
Two set；The similarity weights of the reference data and described problem data are calculated according to the depth value.Pass through the present embodiment
The similarity of paths analysis method provided, can will be tree-like with the second of the first tree data of reference data and problem data
Each node of data is indicated by way of set, and each set includes and depth value for representing each node level
Set, and according to the similarity weights of each depth value calculating reference data and described problem data, pass through the present embodiment meter
The similarity weights of the reference data and described problem data obtained by calculating can determine reference data and problem data it
Between corresponding relation, and then utility problem intelligent analysis module can automate performance positioning work, and performance is greatly improved
The time of case study, reduce the dependence to manpower.
Brief description of the drawings
A kind of preferred embodiment steps flow chart for the similarity of paths analysis method that Fig. 1 is provided by the embodiment of the present invention
Figure；
Another preferred embodiment steps flow chart for the similarity of paths analysis method that Fig. 2 is provided by the embodiment of the present invention
Figure；
A kind of preferred embodiment structure of first tree data of the reference data that Fig. 3 is provided by the embodiment of the present invention is shown
It is intended to；
A kind of preferred embodiment structural representation for the system that Fig. 4 is provided by the embodiment of the present invention；
Another preferred embodiment structural representation for the system that Fig. 5 is provided by the embodiment of the present invention；
Another preferred embodiment structural representation for the system that Fig. 6 is provided by the embodiment of the present invention.
Embodiment
The embodiments of the invention provide a kind of similarity of paths analysis method, the corresponding relation of the mutation for solving path
Problem, thus realize automation performance positioning work, greatly improve performance issue analysis time, reduce to manpower according to
Rely.
Below in conjunction with the similarity of paths analysis method of offer is described in detail the present embodiment shown in Fig. 1：
As shown in figure 1, the similarity of paths analysis method includes：
101st, the first tree data of reference data and the second tree data of problem data are obtained respectively；
Data i.e. to input are parsed, to obtain first tree data and problem data of reference data respectively
Second tree data；
Wherein, the data preferably inputted are I/O path, and certain the present embodiment is only illustrated with I/O path citing, no
It is construed as limiting, the data of such as input can also be the request paths in distributed file system.
Specifically, the canonical form of the I/O path of input can be the dot texts of xml tree structures or graphviz softwares
Part form.
102nd, according to the membership between identical traversal rule and each node to first tree data and institute
Each node for stating the second tree data is traveled through to obtain the depth value of each node；
Wherein, the depth value of each node is used to represent each node in first tree data or the second tree data
Middle be located at level；
Layer of each node in first tree data or second tree data is represented by the depth value
It is secondary.
The present embodiment is not construed as limiting to the traversal rule, as long as to first tree data and described second tree-like
Data are traveled through according to identical traversal rule, and according to the traversal rule can to first tree data and
All nodes in second tree data are traveled through to obtain the depth value of all nodes.
103rd, is determined according to the depth value of first tree data and each node of second tree data respectively
One set and second set；
Wherein, the first set includes multiple first subsets corresponding with the depth value respectively, and each described first
The depth value of the node of first tree data in subset is identical；
The second set includes multiple yield in the second subset corresponding with the depth value respectively, and in each yield in the second subset
Second tree data node depth value it is identical；
I.e. by the step 103 shown in the present embodiment by with each node of the reference data of tree represenation and problem data
It is indicated by way of set.
104th, according to each first subset and depth value corresponding with each yield in the second subset calculates the base value
According to the similarity weights with described problem data.
Pass through the similarity weights of corresponding with each node depth value calculating benchmark data and problem data respectively, Jin Erying
Corresponding relation algorithm is mutated with I/O path and determines that final path is mutated relation.
Specifically, the similarity analysis of the similarity weights i.e. to calculate the reference data and described problem data is calculated
Method is calculated the node of different depth values based on each node being indicated in the way of set, finally exports two classes
The similarity weights of I/O path data.
The similarity of paths analysis method provided by the present embodiment, can be by with the first tree data of reference data
Be indicated with each node of the second tree data of problem data by way of set, and each set include with for representing
The depth value subclass of each node level, and according to each depth value calculating reference data and the similarity of described problem data
Weights, the similarity weights of the reference data and described problem data obtained by being calculated by the present embodiment can determine base
Corresponding relation between quasi- data and problem data, and then utility problem intelligent analysis module can automate performance positioning
Work, is greatly improved the time of performance issue analysis, reduces the dependence to manpower.
Below in conjunction with the nodes of different levels is pointed to shown in Fig. 2 to that can realize with different algorithms to obtain similarity
Weights are described in detail：
201st, the first tree data of reference data and the second tree data of problem data are obtained respectively；
Specifically step 101 please as shown in Figure 1, is not repeated in the present embodiment.
202nd, according to the membership between identical traversal rule and each node to first tree data and institute
Each node for stating the second tree data is traveled through to obtain the depth value of each node；
Wherein, the depth value of each node is used to represent each node in first tree data or the second tree data
Middle be located at level；
The traversal rule is illustrated the present embodiment, need it is clear that, the traversal rule shown in the present embodiment
It is only a kind of example, is not construed as limiting, as long as to first tree data and second tree data according to identical
Traversal rule is traveled through, and can be to first tree data and the second tree-like number according to the traversal rule
All nodes in are traveled through to obtain the depth value of all nodes.
Below as shown in figure 3, Fig. 3 is the example of the first tree data of reference data got, need it is clear that,
Fig. 3 show the citing of tree data, is not construed as limiting, by the way that only traversal rule is illustrated shown in Fig. 3.
It is to the process that the first tree data shown in Fig. 3 is traveled through specifically：
1) root node, depth level=1 residing for record node, are read；
Wherein, when reading the first tree data shown in Fig. 3, it may be determined that node A is root node, and by node A depth
Value is recorded as 1；
Judge whether also have child node under root node；
If 2), comprising child node, it is determined that the child node is first object node, and determine sub under the first object node
The number of node, if the number of child node is 1 under first object node, first object node is saved with son under first object node
The depth value of point is equal；If the number of child node is more than the depth of child node under 1, first object node under first object node
It is worth and plus 1 for first object node depth value, if at least one brother of the first object node simultaneously with the first object node
Child node of younger brother's node simultaneously and under the first object node is connected, then the depth of the child node under the first object node
The depth value being worth for the first object node subtracts one；
3) node depth value level, the sub- I/O path of Recursion process are recorded, until determine to traverse the second destination node, should
Second destination node is terminal node；
Wherein, the principle of record node depth value can be：
As shown in Figure 3, i.e., under decision node A whether also have child node, it is seen then that have under node A three child nodes (node B,
Node C and node D)；
As shown in Figure 3, node A child node B is initially directed to by depth-first principle, and records child node B depth value
Level=level+1=2；
Traversal rule shown in the present embodiment can combine the principle of depth-first and breadth First, i.e., shown in traversing graph 3
The order of the first tree data be first order from left to right；
The depth value for determining node B is that after 2, according to depth-first principle, the sub- I/O path of Recursion process determines first
Whether the number of node B child node is more than 1, and traverse node B child node E, and record child node E depth value level
=level+1=3；
Node E child node number is determined, node E child node K is determined according to depth-first principle, and records node K
Depth value level=level+1=4；
The child node for determining node K is N, because node N father node has multiple, then node N depth value is node K depth
Angle value subtracts one, i.e. record and the mono- N of Ei depth value level=level-1=3；
The child node for determining node N is node S, and node S is terminal node, and because node S father node has multiple, then
It is the second destination node to determine node S, and records child node S depth value level=3-1=2；
4), determine the second destination node N whether be the first tree data global terminal, if it is not, then by second target
The depth value press-in data storage stack of node, and current subpath stacked data is preserved, the root node of the subpath is pointed to, is carried out
Another IO subpaths traversal, until traversing the last item IO subpaths；
Wherein, whether specifically determine second destination node is that the concrete mode of global terminal of the first tree data can
To judge whether also have child node under second destination node；
In shown in Fig. 3, there is child node Q under node S, you can it is not the global terminal of the first tree data to determine node S,
But the terminal node of subpath that node S is located at；
Node S depth value 2 is pressed into data storage stack, and preserves the subpath stacked data traveled through, i.e., successively will
Node N depth value 3, node K depth value 4, node E depth value 3, node B depth value 2 and node A depth value 1
It is pressed into data storage stack；
The root node A of the node S subpath being located at is pointed to, according to breadth First principle, another strip IO roads are carried out
The traversal in footpath, then traverse node F, record node F depth value level=level+1=via node A and node B successively
3；
Because of node F child node only one of which, then depth value is constant, then node F and node P is in the first tree data
Same level, then it is 3 that node F depth value and node P depth value are equal；
Node F child node is node S, and node S has traveled through completion, then successively by node P depth value 3 and node F
Depth value 3 press-in data storage stack in；
The root node A of the node P subpath being located at is pointed to, according to breadth First principle, another strip IO roads are carried out
The traversal in footpath, then traverse node A child node C via node A, and records node C depth value level=level+1=
2；
It can determine that node G and node C is located at same level according to above-mentioned traversal rule, then node C and node G depth value
It is 2, and node G is the second destination node, and node G is not global terminal, and successively by node G and node C depth value
It is pressed into data storage stack；
According to breadth First principle, the traversal of another strip I/O path is carried out, then node A is traversed via node A
Node D, and record node D depth value level=level+1=2；
Node H depth value level=level+1=3 is can determine that according to above-mentioned traversal rule；
Node H and node M are located at same level, then node H and the depth value of node M are 3, and node M is the second mesh
Mark node；
The terminal node for determining the single sub path is node O, because there is child node Q under node O, it is determined that node O is not
Global terminal, and because node O has multiple father nodes (node M, node I and node J) connection, then node O depth value level
There are multiple father nodes on=level-1=2, even child node, then the depth value of child node is that father node depth value subtracts one；
Then record node O depth value 2, and node O, node M, node H and node D depth value are pressed into number successively
According to stack；
According to breadth First principle, the traversal of another I/O path is carried out, i.e., traverses section via node A, node D successively
Point I, from above-mentioned traversal rule, node I depth value level=level+1=3；
Node I child node is node O, because node O depth value has been pressed into data storage stack, then by node I depth
Value press-in data storage stack；
If 5) traverse the last item subpath, global terminal is traversed successively, and by the depth value pressure of global terminal
Enter data storage stack；
According to above-mentioned traversal rule, it is determined that it is the last item subpath to traverse the subpath that node J is located at, then according to
Depth value level=level-1=1 of the node J depth value for 3, node Q is determined successively according to above-mentioned traversal rule, then successively
Node Q depth value 1 and node J depth value 3 are pressed into data storage stack successively；
The depth value of each node of the first tree data, the second tree-like number can be obtained according to above-mentioned traversal rule
According to each node depth value acquisition modes see the depth value of each node of the first tree data acquisition modes, specifically not
Repeat again, need it is clear that, as long as first tree data and second tree data enter according to identical traversal rule
Row traversal.
203rd, is determined according to the depth value of first tree data and each node of second tree data respectively
One set and second set；
By being carried out step 202 Suo Shi to first tree data in ergodic process, by taking first tree data as an example,
The depth value of each node is pressed into data storage stack successively, the sequencing in press-in data storage stack sets up each first
Subset；
From the foregoing, first subset corresponding with depth value 1 is：{ node A, node Q }；
First subset corresponding with depth value 2 is：{ node S, node B, node G, node C, node O, node D }；
First subset corresponding with depth value 3 is：Node N, node E, node P, node F, node M, node H, node I,
Node J }；
First subset corresponding with depth value 4 is：{ node K, node L }；
It can be seen that, the first set includes multiple first subsets corresponding with the depth value respectively, and each described first
The depth value of the node of first tree data in subset is identical；
The specific mode of setting up of the second set is asked for leave the mode of setting up of first set, and is not done in the present embodiment specifically
Repeat, as long as the second set includes multiple yield in the second subset corresponding with the depth value respectively, and each yield in the second subset
The depth value of the node of interior second tree data is identical.
204th, the maximum and determination and the yield in the second subset pair of each depth value corresponding with first subset are determined
The maximum for each depth value answered；
205th, determine corresponding with first subset each depth value maximum and it is corresponding with the yield in the second subset respectively
Smaller value in the maximum of depth value is the first reference depth value；
206th, determine corresponding with first subset each depth value maximum and it is corresponding with the yield in the second subset respectively
Higher value in the maximum of depth value is the second reference depth value；
There is the first set shown in above-mentioned steps to understand, each depth corresponding with each first subset in the first set
The maximum of angle value is 4；Assuming that determining each depth value corresponding with each yield in the second subset in second set by identical traversal rule
It is maximum for 10, it is determined that the first reference depth value is 4, and the second reference depth value is 10.
207th, determine that initialization similarity weights W is equal to 0；
Wherein, W represents the similarity weights of reference data and problem data, and the smaller similarities that represent of W are smaller, and 0 represents two
Class data are entirely different.
208th, it is c to determine destination node number；
Wherein, the c is equal to each first subset and the maximum of the nodes in each yield in the second subset.
It is maximum with the nodes in first subset corresponding with depth value 3 in first set in the present embodiment, then can be true
It is 8 to determine c values.
Need it is clear that, the precedence relationship in sequential is had no between step 204 to the step 208 in the present embodiment, it enters
Capable order is only a kind of citing in the present embodiment.
And setting of the present embodiment to c is also a kind of citing, is not construed as limiting, in different application scenarios and to similarity
Under difference is required, c values can be arbitrarily set, numerical value, i.e. the present embodiment such as also c values can be set as to 500,1000 are set to c
Surely it is not construed as limiting.
209th, according to each first subset and depth value corresponding with each yield in the second subset order from low to high
Determine whether current depth value is less than or equal to first reference depth value successively, if so, step 210 is then carried out, if it is not, then
Carry out step 214；
Each subset in first set and second set is traveled through respectively, and the order of traversal is by low by depth value
To high order, i.e., by taking the first set of above-mentioned determination as an example, first determine the depth value of the first subset { node A, node Q } to work as
Preceding depth value, i.e. current depth value are 1；It is current depth value currently to travel through the depth value analyzed.
Determine whether current depth value 1 is less than or equal to first reference depth value 4,；
Determine after the first subset { node A, node Q }, the first subset { section is determined again by the order of depth value from low to high
Point S, node B, node G, node C, node O, node D }, it is then determined that { node N, node E, node P, node F, node M, section
Point H, node I, node J }, until determining to arrive last first subset { node K, node L }；
The traversal mode of second set is identical with first set, i.e., according to depth value corresponding with each yield in the second subset by
Low to high order determines whether current depth value is less than or equal to first reference depth value successively, specifically repeats no more.
210th, corresponding with the current depth value the first subset and described the are determined according to string editing distance algorithm
SED values between two subsets；
Wherein, string editing distance algorithm (SED, string edit distance), refers between two character strings, by
One minimum edit operation number of times for being converted into needed for another, the edit operation of license includes a character being substituted for another
Individual character, inserts a character, deletes three kinds of a character.SED can be with the difference of two character strings of quantitative measurement；For example will
Kitten changes into sitting:Specific conversion process can carry out sitten (k- first>S), sittin (e- are carried out afterwards>I), most
Afterwards carry out sitting (->G), it is seen that SED=3；I.e. SED distances are used for example in nature frequently as a kind of Similarity Measure function
In Language Processing, spell check, webpage similarity comparison, it is that a kind of effective weigh is marked to apply on I/O path similarity analysis
Standard, specifically see shown in prior art, is not repeated specifically in the present embodiment.
When determining that current depth value is equal to 1, the SED values between corresponding with depth value 1 first subset and yield in the second subset,
The specific result for calculating SED values according to the character string in first subset corresponding from depth value 1 and yield in the second subset it is different without
Together, then do not repeat in the present embodiment, specifically how to calculate SED values according to node is prior art, is repeated no more.
211st, judge whether SED values corresponding with the current depth value are equal to 0, if so, step 212 is then carried out, if it is not,
Then carry out step 213；
212nd, the similarity weights W=W+a^ is determined(b-1)*2c；
Wherein, a span is that concrete numerical value is not construed as limiting in the present embodiment, for example, can be more than 0 and less than 1
1/2；
B is the current depth value, and b is more than or equal to 1 and less than or equal to first reference depth value.
213rd, the similarity weights W=W+a^ is determined(b-1)*2c/SED；
Depth value is traveled through one time from 1 to first reference depth value via above-mentioned steps, and then exports a phase
Like degree weights, following step is carried out on the basis of the similarity weights of output；
214th, determine whether the current depth value is less than or equal to second reference depth value, if so, then being walked
Rapid 215, if it is not, then carrying out step 216；
The second reference depth value is illustrated exemplified by 10 described in the present embodiment, i.e., be more than the first base in current depth value
During quasi- depth value, then determine whether whether current depth value is less than or equal to second reference depth value.
215th, the similarity weights W=W-a^ are determined(b-1)；
Wherein, b is more than first reference depth value and less than or equal to second reference depth value.
216th, terminate flow, and the similarity weights W finally determined is exported；
As illustrated in the present embodiment so that second reference depth value is 10 as an example, then when current depth value is 10
When, you can it will be that 10 corresponding W are exported with depth value, and terminate flow.
Shown in the present embodiment, because with depth value from 1 to the corresponding each node of first reference depth value be
In reference data and problem data be upper layer data, depth value from more than first reference depth value to less than or equal to the second base
The corresponding each node of quasi- depth value is lower data in reference data and problem data, because upper side data and lower data are to base
Between quasi- data and problem data the influence of corresponding relation be it is different, so the present embodiment in calculating and depth value from 1 to institute
When stating the similarity weights of the corresponding each node of the first reference depth value, to be pointed to different layers in data by c setting
Secondary node uses different calculations, so that the node of different levels assigns different power when calculating similarity weights
Value, so as to improve the identification certainty of similarity, has ensured the accurate judgement of structural mutation corresponding relation, so as to improve point
Analyse the accuracy of result.
It is further known that, can be by different reference datas by the similarity of paths analysis method shown in the present embodiment
It is mapped with problem data, the blank filled up between Trace systems and performance issue analysis tool effectively helps product
The positioning work of performance issue.
Carried out below in conjunction with a kind of system for realizing route similarity analysis method that Fig. 4 is provided the present embodiment
Describe in detail：
As shown in figure 4, the system includes：
First acquisition unit 401, for obtaining the first tree data of reference data and the second tree of problem data respectively
Graphic data；
Second acquisition unit 402, for according to the membership between identical traversal rule and each node to described
Each node of first tree data and second tree data is traveled through to obtain the depth value of each node, the institute of each node
Stating depth value is used to represent the level that each node is located in first tree data or the second tree data；
First determining unit 403, for respectively according to first tree data and each section of second tree data
Point depth value determine first set and second set, wherein, the first set include it is multiple respectively with the depth value pair
The depth value of the node of first tree data in the first subset answered, and each first subset is identical, and described second
Set includes multiple yield in the second subset corresponding with the depth value respectively, and the described second tree-like number in each yield in the second subset
According to node depth value it is identical；
Computing unit 404, by according to each first subset and based on depth value corresponding with each yield in the second subset
Calculate the similarity weights of the reference data and described problem data.
The system provided by the present embodiment, can be by with the of the first tree data of reference data and problem data
Each node of two tree datas is indicated by way of set, and each set includes and the depth for representing each node level
Angle value subclass, and according to the similarity weights of each depth value calculating reference data and described problem data, by this reality
The similarity weights for applying the reference data obtained by example is calculated and described problem data can determine reference data and problem
Corresponding relation between data, and then utility problem intelligent analysis module can automate performance positioning work, greatly carry
The time of high-performance question analysis, reduce the dependence to manpower.
Below in conjunction with the nodes of different levels is pointed to shown in Fig. 5 to that can realize with different algorithms to obtain similarity
The system of weights continues to describe in detail：
As shown in figure 5, the system includes：
First acquisition unit 501, for obtaining the first tree data of reference data and the second tree of problem data respectively
Graphic data；
Second acquisition unit 502, for according to the membership between identical traversal rule and each node to described
Each node of first tree data and second tree data is traveled through to obtain the depth value of each node, the institute of each node
Stating depth value is used to represent the level that each node is located in first tree data or the second tree data；
First determining unit 503, for respectively according to first tree data and each section of second tree data
Point depth value determine first set and second set, wherein, the first set include it is multiple respectively with the depth value pair
The depth value of the node of first tree data in the first subset answered, and each first subset is identical, and described second
Set includes multiple yield in the second subset corresponding with the depth value respectively, and the described second tree-like number in each yield in the second subset
According to node depth value it is identical；
Second determining unit 504, maximum and determination for determining each depth value corresponding with first subset
The maximum of each depth value corresponding with the yield in the second subset；
3rd determining unit 505, for determine corresponding with first subset each depth value maximum and with it is described
Smaller value in the maximum of the corresponding each depth value of yield in the second subset is the first reference depth value；
4th determining unit 506, for determine corresponding with first subset each depth value maximum and with it is described
Higher value in the maximum of the corresponding each depth value of yield in the second subset is the second reference depth value；
5th determining unit 507, for determining that initialization similarity weights W is equal to 0；
6th determining unit 508, for determining that destination node number is c, wherein, the c is equal to each first subset
With the maximum of the nodes in each yield in the second subset；
Computing unit 509, by according to each first subset and based on depth value corresponding with each yield in the second subset
Calculate the similarity weights of the reference data and described problem data；
Specifically, the computing unit 509 includes：
First determining module 5091, for according to each first subset and depth corresponding with each yield in the second subset
The order of value from low to high determines whether current depth value is less than or equal to first reference depth value successively；
First computing module 5092, if being less than or equal to first reference depth value, basis for current depth value
String editing distance algorithm determines the SED between the first subset corresponding with the current depth value and the yield in the second subset
Value；
Second computing module 5093, if being equal to 0 for the SED values, it is determined that the similarity weights W=W+a^(b-1)*
2c, wherein, a span is more than 0 and is the current depth value less than 1, b, and b is more than or equal to 1 and is less than or waits
In first reference depth value；
3rd computing module 5094, if being not equal to 0 for the SED values, it is determined that the similarity weights W=W+a
^(b-1)*2c/SED。
Second determining module 5095, if for determining current depth value successively according to the order of the depth value from low to high
During more than first reference depth value, it is determined that whether the current depth value is less than or equal to second reference depth
Value；
3rd determining module 5096, for if it is determined that the current depth value is less than or equal to second reference depth
Value, it is determined that the similarity weights W=W-a^(b-1), wherein, b is more than first reference depth value and less than or equal to institute
State the second reference depth value.
Shown in the present embodiment, because with depth value from 1 to the corresponding each node of first reference depth value be
In reference data and problem data be upper layer data, depth value from more than first reference depth value to less than or equal to the second base
The corresponding each node of quasi- depth value is lower data in reference data and problem data, because upper side data and lower data are to base
Between quasi- data and problem data the influence of corresponding relation be it is different, so the present embodiment in calculating and depth value from 1 to institute
When stating the similarity weights of the corresponding each node of the first reference depth value, to be pointed to different layers in data by c setting
Secondary node uses different calculations, so that the node of different levels assigns different power when calculating similarity weights
Value, so as to improve the identification certainty of similarity, has ensured the accurate judgement of structural mutation corresponding relation, so as to improve point
Analyse the accuracy of result.
It is further known that, can be by different reference datas by the similarity of paths analysis method shown in the present embodiment
It is mapped with problem data, the blank filled up between Trace systems and performance issue analysis tool effectively helps product
The positioning work of performance issue.
The structure of system is described in detail the angle of embodiment slave module functional entity shown in Fig. 4 to Fig. 5, with
The system in the embodiment of the present invention is described in detail from hardware point of view by lower combination Fig. 6, see Fig. 6, in the embodiment of the present invention
Another embodiment of system include：
The system 600 is specifically included：
Input unit 601, output device 602, processor 603 and (wherein, the processor 603 shown in Fig. 6 of memory 604
Can have one or more, be illustrated in Fig. 6 by taking a processor 603 as an example)；
In some embodiments of the invention, input unit 601, output device 602, processor 603 and memory 604 can lead to
Bus or other manner connection are crossed, wherein, in Fig. 6 exemplified by being connected by bus.
Processor 603 is used to perform following steps：
For obtaining the first tree data of reference data and the second tree data of problem data respectively；
For according to the membership between identical traversal rule and each node to first tree data and institute
Each node for stating the second tree data is traveled through to obtain the depth value of each node, and the depth value of each node is used to represent
The level that each node is located in first tree data or the second tree data；
For determining according to the depth value of first tree data and each node of second tree data respectively
One set and second set, wherein, the first set includes multiple first subsets corresponding with the depth value respectively, and respectively
The depth value of the node of first tree data in first subset is identical, the second set include it is multiple respectively with
The corresponding yield in the second subset of the depth value, and the depth value phase of the node of second tree data in each yield in the second subset
Together；
For according to each first subset and depth value corresponding with each yield in the second subset calculates the base value
According to the similarity weights with described problem data.
In other embodiments of the present invention, the processor 603 is additionally operable to perform following steps：
Maximum and determination and the yield in the second subset pair for determining each depth value corresponding with first subset
The maximum for each depth value answered；
For determine corresponding with first subset each depth value maximum and it is corresponding with the yield in the second subset respectively
Smaller value in the maximum of depth value is the first reference depth value；
For determine corresponding with first subset each depth value maximum and it is corresponding with the yield in the second subset respectively
Higher value in the maximum of depth value is the second reference depth value；
For determining that initialization similarity weights W is equal to 0；
For determining that destination node number is c, wherein, the c is equal to each first subset and each yield in the second subset
The maximum of interior nodes.
In other embodiments of the present invention, the processor 603 is additionally operable to perform following steps：
For according to each first subset and depth value corresponding with each yield in the second subset order from low to high
Determine whether current depth value is less than or equal to first reference depth value successively；
If being less than or equal to first reference depth value for current depth value, according to string editing distance algorithm
It is determined that the SED values between the first subset corresponding with the current depth value and the yield in the second subset；
If being equal to 0 for the SED values, it is determined that the similarity weights W=W+a^(b-1)* 2c, wherein, a value model
Enclose to be the current depth value more than 0 and less than 1, b, and b is more than or equal to 1 and less than or equal to first reference depth
Value；
If being not equal to 0 for the SED values, it is determined that the similarity weights W=W+a^(b-1)*2c/SED。
In other embodiments of the present invention, the processor 603 is additionally operable to perform following steps：
If for determining that current depth value is more than first benchmark successively according to the order of the depth value from low to high
During depth value, it is determined that whether the current depth value is less than or equal to second reference depth value；
For if it is determined that the current depth value is less than or equal to second reference depth value, it is determined that the similarity
Weights W=W-a^(b-1), wherein, b is more than first reference depth value and less than or equal to second reference depth value.
It is apparent to those skilled in the art that, for convenience and simplicity of description, the system of foregoing description,
The specific work process of device and unit, may be referred to the corresponding process in preceding method embodiment, will not be repeated here.
In several embodiments provided herein, it should be understood that disclosed system, apparatus and method can be with
Realize by another way.For example, device embodiment described above is only schematical, for example, the unit
Divide, only a kind of division of logic function there can be other dividing mode when actually realizing, such as multiple units or component
Another system can be combined or be desirably integrated into, or some features can be ignored, or do not perform.It is another, it is shown or
The coupling each other discussed or direct-coupling or communication connection can be the indirect couplings of device or unit by some interfaces
Close or communicate to connect, can be electrical, machinery or other forms.
The unit illustrated as separating component can be or may not be it is physically separate, it is aobvious as unit
The part shown can be or may not be physical location, you can with positioned at a place, or can also be distributed to multiple
On NE.Some or all of unit therein can be selected to realize the mesh of this embodiment scheme according to the actual needs
's.
In addition, each functional unit in each embodiment of the invention can be integrated in a processing unit, can also
That unit is individually physically present, can also two or more units it is integrated in a unit.Above-mentioned integrated list
Member can both be realized in the form of hardware, it would however also be possible to employ the form of SFU software functional unit is realized.
If the integrated unit is realized using in the form of SFU software functional unit and as independent production marketing or used
When, it can be stored in a computer read/write memory medium.Understood based on such, technical scheme is substantially
The part contributed in other words to prior art or all or part of the technical scheme can be in the form of software products
Embody, the computer software product is stored in a storage medium, including some instructions are to cause a computer
Equipment (can be personal computer, server, or network equipment etc.) performs the complete of each embodiment methods described of the invention
Portion or part steps.And foregoing storage medium includes：USB flash disk, mobile hard disk, read-only storage (ROM, Read-Only
Memory), random access memory (RAM, Random Access Memory), magnetic disc or CD etc. are various can store journey
The medium of sequence code.
Described above, the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although with reference to before
Embodiment is stated the present invention is described in detail, it will be understood by those within the art that：It still can be to preceding
State the technical scheme described in each embodiment to modify, or equivalent substitution is carried out to which part technical characteristic；And these
Modification is replaced, and the essence of appropriate technical solution is departed from the spirit and scope of various embodiments of the present invention technical scheme.
Claims (8)
1. a kind of similarity of paths analysis method, it is characterised in that including：
The first tree data of reference data and the second tree data of problem data are obtained respectively；
According to the membership between identical traversal rule and each node to first tree data and second tree
Each node of graphic data is traveled through to obtain the depth value of each node, and the depth value of each node is used to represent that each node exists
The level being located in first tree data or the second tree data；
Respectively according to the depth value of first tree data and each node of second tree data determine first set and
Second set, wherein, the first set includes multiple first subsets corresponding with the depth value respectively, and each described first
The depth value of the node of first tree data in subset is identical, the second set include it is multiple respectively with the depth
The depth value for being worth the node of second tree data in corresponding yield in the second subset, and each yield in the second subset is identical；
According to each first subset and depth value corresponding with each yield in the second subset calculates the reference data and described
The similarity weights of problem data.
2. similarity of paths analysis method according to claim 1, it is characterised in that the basis and each first son
Collection and depth value corresponding with each yield in the second subset calculate the reference data and described problem data similarity weights it
Before, methods described also includes：
It is determined that the maximum and determination each depth corresponding with the yield in the second subset of each depth value corresponding with first subset
The maximum of angle value；
It is determined that the maximum and each depth value corresponding with the yield in the second subset of corresponding with first subset each depth value
Smaller value in maximum is the first reference depth value；
It is determined that the maximum and each depth value corresponding with the yield in the second subset of corresponding with first subset each depth value
Higher value in maximum is the second reference depth value；
Determine that similarity weights W is equal to 0；
It is c to determine destination node number, wherein, the c is equal to the node in each first subset and each yield in the second subset
Several maximums.
3. similarity of paths analysis method according to claim 2, it is characterised in that the basis and each first son
Collection and the similarity weights bag with each the yield in the second subset corresponding depth value calculating reference data and described problem data
Include：
Determined successively with the order of depth value corresponding with each yield in the second subset from low to high according to each first subset
Whether current depth value is less than or equal to first reference depth value；
If so, then determining the first subset corresponding with the current depth value and described second according to string editing distance algorithm
SED values between subset；
If the SED values are equal to 0, it is determined that the similarity weights W=W+a^(b-1)* 2c, wherein, a span be more than
0 and be the current depth value less than 1, b, and b is more than or equal to 1 and less than or equal to first reference depth value；
If the SED values are not equal to 0, it is determined that the similarity weights W=W+a^(b-1)*2c/SED。
4. similarity of paths analysis method according to claim 3, it is characterised in that the basis and each first son
Collection and depth value corresponding with each yield in the second subset calculate the similarity weights of the reference data and described problem data also
Including：
If determine that current depth value is more than first reference depth value successively according to the order of the depth value from low to high,
Then determine whether the current depth value is less than or equal to second reference depth value；
If, it is determined that the similarity weights W=W-a^(b-1), wherein, b be more than first reference depth value and be less than or
Equal to second reference depth value.
5. a kind of similarity of paths analysis system, it is characterised in that including：
First acquisition unit, for obtaining the first tree data of reference data and the second tree data of problem data respectively；
Second acquisition unit, for tree-like to described first according to the membership between identical traversal rule and each node
Each node of data and second tree data is traveled through to obtain the depth value of each node, the depth value of each node
For representing the level that each node is located in first tree data or the second tree data；
First determining unit, for respectively according to the depth of first tree data and each node of second tree data
Value determines first set and second set, wherein, the first set includes multiple respectively corresponding with the depth value first
The depth value of the node of first tree data in subset, and each first subset is identical, and the second set includes
Multiple yield in the second subset corresponding with the depth value respectively, and the node of second tree data in each yield in the second subset
Depth value it is identical；
Computing unit, for according to each first subset and depth value corresponding with each yield in the second subset calculates the base
The similarity weights of quasi- data and described problem data.
6. system according to claim 5, it is characterised in that the system also includes：
Second determining unit, maximum and determination and described for determining corresponding with first subset each depth value
The maximum of the corresponding each depth value of two subsets；
3rd determining unit, for determine corresponding with first subset each depth value maximum and with the yield in the second subset
Smaller value in the maximum of corresponding each depth value is the first reference depth value；
4th determining unit, for determine corresponding with first subset each depth value maximum and with the yield in the second subset
Higher value in the maximum of corresponding each depth value is the second reference depth value；
5th determining unit, for determining that similarity weights W is equal to 0；
6th determining unit, for determining that destination node number is c, wherein, the c is equal to each first subset and each described
The maximum of nodes in yield in the second subset.
7. system according to claim 6, it is characterised in that the computing unit includes：
First determining module, for according to each first subset and depth value corresponding with each yield in the second subset by it is low to
High order determines whether current depth value is less than or equal to first reference depth value successively；
First computing module, if being less than or equal to first reference depth value for current depth value, according to character Series Code
Collect the SED values between distance algorithm determination the first subset corresponding with the current depth value and the yield in the second subset；
Second computing module, if being equal to 0 for the SED values, it is determined that the similarity weights W=W+a^(b-1)* 2c, wherein,
A span is more than 0 and be the current depth value less than 1, b, and b is more than or equal to 1 and is less than or equal to described the
One reference depth value；
3rd computing module, if being not equal to 0 for the SED values, it is determined that the similarity weights W=W+a^(b-1)*2c/
SED。
8. system according to claim 7, it is characterised in that the computing unit also includes：
Second determining module, if described for determining that current depth value is more than successively according to the order of the depth value from low to high
During the first reference depth value, it is determined that whether the current depth value is less than or equal to second reference depth value；
3rd determining module, for if it is determined that the current depth value is less than or equal to second reference depth value, it is determined that
The similarity weights W=W-a^(b-1), wherein, b is more than first reference depth value and less than or equal to second benchmark
Depth value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410705903.XA CN104462288B (en) | 2014-11-27 | 2014-11-27 | A kind of similarity of paths analysis method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410705903.XA CN104462288B (en) | 2014-11-27 | 2014-11-27 | A kind of similarity of paths analysis method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104462288A CN104462288A (en) | 2015-03-25 |
| CN104462288B true CN104462288B (en) | 2017-10-17 |
Family
ID=52908324
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410705903.XA Active CN104462288B (en) | 2014-11-27 | 2014-11-27 | A kind of similarity of paths analysis method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104462288B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105184387A (en) * | 2015-07-23 | 2015-12-23 | 北京理工大学 | Path similarity comparison method |
| CN109885733B (en) * | 2019-01-18 | 2020-09-15 | 清华大学 | Graph data compression method and device for target spanning tree query |
| CN112994990B (en) * | 2021-05-20 | 2021-07-30 | 蚂蚁金服（杭州）网络技术有限公司 | Loop detection method and device, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101604365A (en) * | 2009-07-10 | 2009-12-16 | 珠海金山软件股份有限公司 | Determine the system and method for number of computer rogue program sample families |
| CN103823883A (en) * | 2014-03-06 | 2014-05-28 | 焦点科技股份有限公司 | Analysis method and system for website user access path |
| CN104063445A (en) * | 2014-06-16 | 2014-09-24 | 百度移信网络技术（北京）有限公司 | Method and system for measuring similarity |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4767127B2 (en) * | 2006-08-10 | 2011-09-07 | 株式会社日立製作所 | File server, computer system, and file prefetching method. |
-
2014
- 2014-11-27 CN CN201410705903.XA patent/CN104462288B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101604365A (en) * | 2009-07-10 | 2009-12-16 | 珠海金山软件股份有限公司 | Determine the system and method for number of computer rogue program sample families |
| CN103823883A (en) * | 2014-03-06 | 2014-05-28 | 焦点科技股份有限公司 | Analysis method and system for website user access path |
| CN104063445A (en) * | 2014-06-16 | 2014-09-24 | 百度移信网络技术（北京）有限公司 | Method and system for measuring similarity |
Non-Patent Citations (1)
| Title |
|---|
| 基于分支路径跟踪的猜测访存数据Cache污染控制技术;刘松鹤等;《计算机应用研究》;20130731;第30卷(第7期);2064-2067 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104462288A (en) | 2015-03-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9679248B2 (en) | Automated creation of semantically-enriched diagnosis models using time series data of temperatures collected by a network of sensors | |
| US7801924B2 (en) | Decision tree construction via frequent predictive itemsets and best attribute splits | |
| CN106156145A (en) | The management method of a kind of address date and device | |
| US9984134B2 (en) | Extraction device, data processing system, and extraction method | |
| CN110377682B (en) | Track type determination method and device, computing equipment and storage medium | |
| CN104462288B (en) | A kind of similarity of paths analysis method and system | |
| CN103425584A (en) | Large-scale application regression testing information processing method based on Java bytecode | |
| CN104834717A (en) | Web information automatic extraction method based on webpage clustering | |
| JP2019169126A (en) | Artificial intelligence-based triple checking method, device, and computer program | |
| CN112000736B (en) | Spatiotemporal trajectory adjoint analysis method and system, electronic device and storage medium | |
| CN108008936A (en) | A kind of data processing method, device and electronic equipment | |
| CN106294128B (en) | A kind of automated testing method and device exporting report data | |
| CN111916144A (en) | Protein classification method based on self-attention neural network and coarsening algorithm | |
| CN111045670B (en) | Method and device for identifying multiplexing relationship between binary code and source code | |
| CN107767155A (en) | A kind of method and system for assessing user's representation data | |
| CN116881430B (en) | Industrial chain identification method and device, electronic equipment and readable storage medium | |
| CN114092729A (en) | Heterogeneous electricity consumption data publishing method based on cluster anonymization and differential privacy protection | |
| CN107273264B (en) | Method and device for tracking key attribute and associated attribute of object instance | |
| US9165020B2 (en) | String substitution apparatus, string substitution method and storage medium | |
| CN106649315A (en) | Method and device for processing path navigation | |
| CN114692871A (en) | Decision tree training method, waybill type identification device, equipment and medium | |
| CN115563186A (en) | User behavior intention output method based on big data analysis and big data system | |
| KR101565715B1 (en) | Apparatus and Method for generating co-occurrent subgraph in directed graphs | |
| CN109086373B (en) | Method for constructing fair link prediction evaluation system | |
| CN103064923B (en) | A kind of OLAP query spacing computational methods based on dimension hierarchy feature |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220829 Address after: No. 1899 Xiyuan Avenue, high tech Zone (West District), Chengdu, Sichuan 610041 Patentee after: Chengdu Huawei Technologies Co.,Ltd. Address before: 518129 Bantian HUAWEI headquarters office building, Longgang District, Guangdong, Shenzhen Patentee before: HUAWEI TECHNOLOGIES Co.,Ltd. |
|
| TR01 | Transfer of patent right |