CN109558410A - Event matches algorithm based on multi-dimensional content in a kind of information distribution system - Google Patents
Event matches algorithm based on multi-dimensional content in a kind of information distribution system Download PDFInfo
- Publication number
- CN109558410A CN109558410A CN201811530822.5A CN201811530822A CN109558410A CN 109558410 A CN109558410 A CN 109558410A CN 201811530822 A CN201811530822 A CN 201811530822A CN 109558410 A CN109558410 A CN 109558410A
- Authority
- CN
- China
- Prior art keywords
- attribute
- gem
- subscription
- event
- tree
- 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
Abstract
The present invention discloses a kind of event matches algorithm GEM-Tree based on multi-dimensional content aiming at the problem that efficiency that the matching of information distribution system quick obtaining is subscribed to is restricted.This patent devises expression semanteme and more enriches polynary expression model;The Representative properties and attribute number for making full use of subscription are screening out the advantage for mismatching and subscribing to, and devise attribute filter table and subscribe to screen out a large amount of mismatch as early as possible, balance the height of GEM-Tree;Codomain partition structure GEM on triangle, will be devised on codomain Interval Maps to two-dimensional surface, improve matching efficiency of the event matches in codomain;The complexity of comprehensive analysis index structure, design it is extensive it is high-dimensional under GEM-Tree;Design and built million grades of subscription platform under various scenes, compared with event matches algorithm (OP-Index, REIN, BE-Tree, AP-INdex) compare, demonstrate the high efficiency and validity of GEM-Tree algorithm.This patent has deeper meaning for improving event matches efficiency in extensive high-dimensional information distribution system.
Description
Technical field
It is field the invention belongs to information distribution, can be efficiently carried out in high-dimensional large-scale information distribution system
Event matches algorithm.For existing event matches algorithm, matching efficiency is low under extensive high-dimensional scene, becomes publication
The problem of bottleneck of ordering system development, has been specifically designed a kind of ability for being based on content, there is very strong attribute to filter, has made to search for
The Fast Match Algorithm GEM-Tree that space reduces rapidly.
Background technique
Information distribution system needs to accommodate and handles that the order of magnitude is big, abundant in content, polynary information.In the information of magnanimity
In, user demand and the matching of multiple information how is rapidly completed, is information distribution system problem to be solved and information
The core technology of dissemination system.
In more attribute publish/subscribe systems of the large size based on content, as pivotal player, it is responsible for from ordering event matches
Read all subscription that match event is searched in database.However, the continuous expansion of system scale, information become increasingly abundant, search
The continuous improvement of complexity, so that the performance of event matches is by stern challenge.Existing matching algorithm is in such a scenario
Matching efficiency is greatly affected.
Existing algorithm can be divided into tree matching and count matches two types.Although existing algorithm is by technology by setting
It counts index structure organizational protection to subscribe to, reaches and effectively screen out mismatch and subscribe to, improve the purpose of matching efficiency, but presently, there are
The shortcomings that it is as follows: 1) larger when subscription, when the attribute dimensions contained are higher, it is impossible to meet fast for existing mating structure
The matched requirement of speed;2) spatial match, count matches algorithm cannot make full use of the characteristics of subscribing to attribute multidimensional, and sufficiently filtering is ordered
It reads;3) tree structure can expend more resources in progress dynamic regulation;4) it cannot dynamically reconfigure to adapt to Physical Network
Change in topology in network.
Summary of the invention
(1) technical problems to be solved
Event matches algorithm in a kind of information distribution system proposed by the present invention based on multi-dimensional content, the skill for needing to solve
Art problem has: how to construct expression semanteme and enriches polynary model more to express related notion;How to design dimension divide and
Single dimension codomain partition mechanism;How index structure and event matches algorithm are designed;How experiment survey is carried out to the performance of algorithm
Examination and comparative evaluation.
(2) technical solution
In order to solve the above technical problems, the invention proposes the events based on multi-dimensional content in a kind of information distribution system
With algorithm, including the following steps:
Step 1: model construction,
Step 2: the design of attribute filter table,
Step 3: the design of GEM structure,
Step 4: GEM-Tree building,
Step 5: the design of GEM-Tree event matches algorithm,
Step 6: GEM-Tree algorithm analysis,
Step 7: the contrast verification of GEM-Tree performance.
Step 1: model construction:
The correlation model in distribution subscription system is constructed, to express semantic subscription more abundant and event information.It closes
Key is expressed as follows:
Step 2: the design of attribute filter table:
It is filtered by attribute, when carrying out subscription store, according to the Representative properties of subscription and the attribute number contained, will be ordered
It reads and is divided into different grid (mesh).When having carried out event and being matched, according to the attribute and attribute number of event, phase is selected
Answer grid, it is intended to subscription will be mismatched and screened out as early as possible, to reduce the number for executing event matches.
(1) key concept is as follows:
Representative properties: it will can most subscribe to and distinguish best attribute with other subscription.
Representative properties priority inquiry table: the list generated after statistics calculating is carried out to the attribute in subscription and event, is
Subscribe to the foundation for choosing Representative properties.
Mesh: according to the Representative properties of subscription and the attribute number contained, determine to subscribe in attribute filter table belonging to
Grid
(2) design principle:
(a) event centainly includes the Representative properties that matching is subscribed to;
(b) attribute number that matching is subscribed to centainly is less than the attribute number of event.
(3) the advantages of attribute filter table:
(a) to the height for subscribing to progress Preliminary division equilibrium GEM-Tree.
(b) a large amount of mismatch of filtering as early as possible is subscribed to, and reduces later period matched number
Step 3: the design of GEM structure:
The codomain of attribute is mapped as a point in plane space by the division that GEM structure carries out codomain, corresponding to be worth
Domain space is mapped as the upper triangle on plane space, and the division of codomain is mapped as the division to plane space.GEM structure is seen appendix
Mapping relations in Fig. 1 .GEM structure chart (1) annex map 1 are as follows:
Three-legged structure in plane: the several structures of plane of Value space mapping.
Section: the point in plane in triangle.
Monodrome: the point in plane on triangle bevel edge.
Cell: the unit obtained after being divided to triangle in plane.
Fig. 2 .GEM carries out the explanation of event matches in attachment
A point | Attribute value |
I1, I2 | Matching area dividing line |
Region A | Matching area |
Region B, C | It mismatches completely |
Region A ', B ' | Region to be matched |
(2) event matches are executed on GEM:
When subscription is inserted into, according to the codomain of attribute, by subscription store in the cell into GEM.When event is come
Face, when carrying out category event matches, from the attribute value value point a of event, draws cut-off rule I1, I2.I1, I2 mark off matching area A,
Mismatch region B and C and region A ' and B ' to be matched.Need to A ' and B ' in subscription carry out further event matches,
Mismatch is screened out to subscribe to.
Step 4: the design of GEM-Tree:
The structure of GEM-Tree such as attached drawing 3.GEM-Tree structure, the node specification in attached drawing 3. are as follows:
The building process of GEM-Tree:
(1) attribute priority list is generated;
(2) Representative properties are added to subscribe to;
(3) according to Representative properties and attribute number, subscription is divided into corresponding mesh;
(4) according to the priority of the attribute of subscription and attribute, further division will be subscribed into corresponding GEM structure
(GEM)
(5) according to the codomain of attribute, subscription is divided into corresponding cell.Check next layer of the GEM table of cell
In in whether containing subscribe in attribute, if so, step 5 is then continued to execute, otherwise by the S- under subscription store to cell
In node;
(6) check whether the subscription in S-node exceeds tolerance after storing.If exceeding, divided in S-node;It is no
Then insertion terminates, and building terminates.
(7) S-node is divided, and is counted to the attribute of subscription, is selected optimum attributes.Generate the GEM of this attribute
Structure extracts the subscription containing this attribute out from S-node, and storage is into newly-built GEM in corresponding cell.
Step 5: the design of GEM-Tree event matches algorithm
The process that GEM-Tree carries out event matches is as follows:
(1) it according to the attribute of event and attribute number, is determined in attri-filter-table and carries out matched mesh;
(2) according to the attribute in event, the corresponding GEM of attribute is selected to be matched;
(3) on GEM structure (upper triangle), according to the value of attribute, cut-off rule delimited in upper triangle, cut-off rule pierced through
Cell carry out into event matches, screen out mismatch and subscribe to.
(4) traversal cell is connected to the list of next layer of GEM structure, if the attribute that discovery event contains, repeats step
2,3,4, until the list of the next layer of GEM structure of the cell at place is not present, or search less than respective attributes.
(5) all matched subscription are combined, obtains final matching result.
Step 6: GEM-Tree analysis of complexity
Relevant parameter is defined as follows:
(1) the predicate number contained in subscribing to is set as λ.It is maximum in subscription if the distribution of predicate is uniformly distributed in subscribing to
Predicate number is λmax, then the average predicate number subscribed to is
(2) size of dimensional space is d in system.
(3) in GEM, the size of the storage space S-node of cell is c.
Subscribe to the time complexity of insertion:
The time complexity for subscribing to insertion is O (λavglogd).It proves as follows.Firstly, being arranged according to Representative properties priority
Table is O (λ to subscribe to the complexity of selection Representative propertiesavg);Second step is the subscription according to its typical attribute and predicate
Selecting the complexity of corresponding grid (mesh) in amount dependence filter table is (1) O;It is to find phase in GEM list in next step
The GEM structure answered, complexity are O (1);Event matches are completed from GEM structure, the complexity for finding corresponding matching codomain is O
(1);In this case, the cost for being inserted into subscription depends on the height of GEM-Tree.The height of GEM-Tree is O (λavgLogd), in conclusion the time complexity for being inserted into subscription in GEM-Tree is O (λavglogd)。
The time complexity of event matches:
The time complexity of event matches is O (λavgc).It proves as follows: finding out matching in dependence filter table and subscribe to place
Grid (mesh) complexity be O (1);The corresponding GEM structure in event containing attribute is found out in next layer of GEM table of grid
Complexity is O (1);The complexity that matching subscription store node S-node is found from GEM-Tree is O (λavglogd);From S-
It is O (λ that the complexity that matching is subscribed to finally is obtained in nodeavgc).In conclusion the time that GEM-Tree completes event matches is multiple
Miscellaneous degree is O (λavgc)。
Step 7: the contrast verification of GEM-Tree performance
(1) experiment porch is built: having been built information capacity and has been subscribed to up to million grades of experiment porch from quantity is subscribed to
Matching degree, the size of attribute space, many factors such as attribute number of subscription consider, design several scenes, to matching algorithm into
Row Experimental comparison and Performance Evaluation.
(2) source is subscribed in experiment: the data packet provided BE-Tree.
(3) it compares algorithm: selecting matching algorithm AP-Tree, BE-Tree, OP- from tree structure and count matches algorithm
Index, REIN, algorithm reproduction is carried out, carries out Experimental comparison with GEM-Tree algorithm.
Experimental comparison results are summarized as follows:
Influence of the size of attribute space for event matches is not very big, but GEM-Tree carries out event matches and has one
A little advantages (attached drawing 4);With the growth for subscribing to quantity, the average event matches time of various algorithms and subscription erasing time are gradually
Increase.For GEM-Tree compared with other algorithms, the speed of growth will slowly very much (5 attached drawings 9 of attached drawing);As what subscription contained is averaged
The growth of attribute number, the time that event matches are spent increase.GEM-Tree shows optimal in this five algorithms.It subscribes to and belongs to
Property number for GEM-Tree carry out event matches influence very little (attached drawing 6);With the growth of subscribing matching rate, event matches flower
The time of expense increases.GEM-Tree shows optimal (attached drawing 5) in this five algorithms.
(3) beneficial effect
(1) it excludes to mismatch as early as possible to subscribe to
All event matches algorithms are by reducing search space, when executing matching to reduce matched event, from
And improve the efficiency of event.Most of existing algorithms pass through the structure of the multiple sub- sections of multilayer, point since a root node
Subscription is divided.However, index structure can be generated accordingly greatly, the child node of amount with the increase for subscribing to the order of magnitude, this dimension
It protects these child nodes to be very expensive, therefore the efficiency of event matches is also affected.In addition, the imbalance of index structure exists
These algorithms are inevitable.GEM-Tree of the present invention replaces root node to ask to alleviate this by using attribute filter table
Topic.Attribute filter table not only contributes to eliminate a large amount of unmatched subscription, additionally aids the height of balance GEM-Tree.
(2) matching of one-dimensional codomain is faster completed
Some existing event matches algorithms, to codomain further division, are subscribed to by layered structure to divide.Event
Subscription space where subscription of the timing by searching candidate matches in hierarchical structure.In contrast, GEM will be worth in the present invention
Domain space is converted into the equilateral upper triangle in a two-dimensional plane, is drawn by what the division to triangle on this was completed to subscribe to
Point.When GEM carries out event matches, the cut-off rule drawn by matching value obtains matching area at once, substantially increases event
The efficiency matched.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be to needed in the embodiment attached
Figure is briefly described, it should be understood that the following drawings illustrates only certain embodiments of the present invention, therefore is not construed as pair
The restriction of range for those of ordinary skill in the art without creative efforts, can also be according to this
A little attached drawings obtain other relevant attached drawings.
Fig. 1 provides subscription insertion flow chart for the present invention
Fig. 2 provides event matches flow chart for the present invention
Fig. 3 provides GEM structure chart for the present invention
Fig. 4 provides GEM event matches figure for the present invention;
Fig. 5 provides GEM-Tree structure chart for the present invention;
Fig. 6 provides influence of the growth to match time of attribute space for the present invention
Fig. 7 is that the present invention proposes influence of the matching rate available on subscription to the event matches time
Fig. 8 is that the present invention proposes influence of the average properties number available on subscription to match time
Fig. 9 provides influence of the growth to match time of subscription for the present invention
Figure 10 provides the influence for subscribing to quantity to the index structure generation time is generated for the present invention
Figure 11 provides the influence for subscribing to quantity to the subscription time is deleted for the present invention
Specific implementation
With reference to the accompanying drawings and examples, to the specific of the building of event matches index structure in the present invention and event matches
Embodiment is described in further detail.The following examples are intended to illustrate the invention, but the range being not intended to limit the invention.
(1) selection of Representative properties
Representative properties are to subscribe to most especially, most characteristic attribute.It can be used as the mark of subscription, being capable of maximum journey
It will subscribe on degree and distinguished with other subscription.The Representative properties of one subscription do not have in other subscription or this frequency of occurrences
It is very small.
According to stream or flow of event is subscribed in this patent, the frequency of each attribute is counted, is ranked up, typical case is ultimately generated
Attribute priority list.When executing the operation that insertion is subscribed to, first according to attribute priority list, type genus is selected to subscribe to
Property (as shown in algorithm one).
(2) generation of attribute filter table
It is filtered by attribute, when carrying out subscription store, according to the Representative properties of subscription and the attribute number contained, will be ordered
It reads and is divided into different grid (mesh).When having carried out event and being matched, according to the attribute and attribute number of event, phase is selected
Grid is answered, subscription will be mismatched and screened out as early as possible, to reduce the number for executing event matches.The principle that attribute filtering follows is: 1)
Event centainly includes the Representative properties that matching is subscribed to;2) attribute number that matching is subscribed to centainly is less than the attribute number of event.Attribute mistake
Filter table has the advantage that: 1) to the height for subscribing to progress Preliminary division equilibrium GEM-Tree;2) a large amount of mismatch of filtering as early as possible is ordered
It reads, reduces later period matched number.
The life algorithm of attribute filter table is as follows:
Predicate quantity in subscription is limited.We are the available maximum predicate number preMaxNum into subscription, with
And system property space attrSet.Attribute filter table is constructed according to preMaxNum, attrSet.Firstly, our computation attributes
The length (line4) of a line in filter table, then traverses all property systems, establishes row attribute and adeditive attribute filter table
Row (line5-11).The density of grid is adjustable in attribute filter table, by predicate increment on specified line direction come
It realizes.
(3) insertion subscribed in GEM-Tree
The invention proposes GEM-Tree, a kind of event matches calculation based in information distribution system based on multi-dimensional content
The building of method, GEM-Tree is as follows, with reference to the structure of attached drawing 3.GEM-Tree:
Step 1: attribute priority list is generated;
Step 2: Representative properties are added to subscribe to;
Step 3: according to Representative properties and attribute number, subscription is divided into corresponding mesh;
Step 4: according to the priority of the attribute of subscription and attribute, further division will be subscribed to and tied to corresponding GEM
In structure (T-node)
Step 5: according to the codomain of attribute, subscription is divided into corresponding cell, and is stored in S-node;
Step 6: check whether the subscription in S-node exceeds tolerance after storage.If exceeding, occur to divide in S-node
It splits;Otherwise insertion terminates, and building terminates.
Step 7: S-node is divided, and is counted to the attribute of subscription, is selected optimum attributes.Generate this attribute
GEM structure, the subscription containing this attribute is extracted out from S-node, storage is into newly-built GEM in corresponding cell.
The specific pseudocode that generates is as shown in algorithm 3.
(4) steps are as follows for the specific execution of event matches:
Step 1: according to the attribute and attribute number of event, determine that progress is matched in attri-filter-table
mesh;
Step 2: according to the attribute in event, the corresponding GEM of attribute is selected to be matched;
Step 3: on GEM structure (upper triangle), according to the value of attribute, cut-off rule delimited in upper triangle, to cut-off rule
The cell of puncture screen out mismatch into event matches and subscribe to.
Step 4: traversal cell is connected to the list of next layer of GEM structure, if the attribute that discovery event contains, repeats
Step 2,3,4 until the list of the next layer of GEM structure of the cell at place is not present, or are searched less than respective attributes.
Step 5: all matched subscription are combined, final matching result is obtained.
The specific pseudocode that generates is as shown in algorithm four:
The present invention is continuously increased for the order of magnitude is subscribed to, and information becomes increasingly abundant, the continuous improvement of search complexity, event
The problem of matching efficiency is greatly affected proposes that the event matches in a kind of information distribution system based on multi-dimensional content are calculated
Method is filtered and codomain by Representative properties, attribute number to mismatching to subscribe to respectively using attribute filter table and GEM structure
Effective division in space, effective acquisition matching are subscribed to, and are greatly improved high-dimensional magnanimity and are subscribed to the efficiency for carrying out event matches.This
Million grades of subscription platform under various scenes is also simulated in invention, builds experiment porch, and compared with other event matches
Algorithm compares, and demonstrates high efficiency of the algorithm of invention in event matches.
It should be understood by those ordinary skilled in the art that: the above is only a specific embodiment of the present invention, and
It is not used in the limitation present invention, all within the spirits and principles of the present invention, any modification, equivalent substitution, improvement and etc. done,
It should be included within protection scope of the present invention.
Claims (7)
1. the event matches algorithm in a kind of information distribution system based on multi-dimensional content, it is characterised in that: S1: constructing model,
The correlation model in distribution subscription system is constructed, to express semantic subscription more abundant and event information;
S2: design attributes filter table is filtered by attribute, when carrying out subscription store, according to the Representative properties of subscription and is contained
Attribute number, be divided into subscribing in different grid (mesh).It is intended to mismatch subscription to screen out as early as possible, be executed with reducing
The number of event matches;
S3: attribute codomain is divided the division being mapped as to plane space by design GEM structure;
S4: design GEM-Tree carries out storage and maintenance to subscribing to;
S5: event matches are carried out according to GEM-Tree;
The contrast verification of S6:GEM-Tree performance.
2. model construction according to claim 1: the correlation model in building distribution subscription system, to express semanteme more
Add subscription abundant and event information.
3. attribute filter table according to claim 1 design: being filtered by attribute, when carrying out subscription store, according to ordering
The Representative properties read and the attribute number contained are divided into subscribing in different grid (mesh).When come event carry out
Timing selects corresponding grid according to the attribute and attribute number of event, it is intended to will mismatch subscription and screen out as early as possible, and be executed with reducing
The number of event matches.
(1) key concept is as follows:
Representative properties: it will can most subscribe to and distinguish best attribute with other subscription.
Representative properties priority inquiry table: carrying out the list generated after statistics calculating to the attribute in subscription and event, is to subscribe to
Choose the foundation of Representative properties.
Mesh: it according to the Representative properties of subscription and the attribute number contained, determines to subscribe to affiliated grid in attribute filter table
(2) design principle:
(a) event centainly includes the Representative properties that matching is subscribed to;
(b) attribute number that matching is subscribed to centainly is less than the attribute number of event.
(3) the advantages of attribute filter table:
(a) to the height for subscribing to progress Preliminary division equilibrium GEM-Tree.
(b) a large amount of mismatch of filtering as early as possible is subscribed to, and reduces later period matched number.
4. the design of GEM structure according to claim 1: the codomain of attribute is reflected in the division that GEM structure carries out codomain
It penetrates as a point in plane space, corresponding Value space is mapped as the upper triangle on plane space, the division mapping of codomain
For the division to plane space.
When subscription is inserted into, according to the codomain of attribute, by subscription store in the cell into GEM.When event arrive, into
It when row belongs to event matches, from the attribute value value point a of event, draws cut-off rule and marks off matching area, mismatch region, and to
Matching area.It needs to treat the subscription in matching area and carries out further event matches, screen out mismatch and subscribe to.
5. the design of GEM-Tree according to claim 1:
The building process of GEM-Tree:
(1) attribute priority list is generated;
(2) Representative properties are added to subscribe to;
(3) according to Representative properties and attribute number, subscription is divided into corresponding mesh;
(4) according to the priority of the attribute of subscription and attribute, further division will be subscribed into corresponding GEM structure (GEM)
(5) according to the codomain of attribute, subscription is divided into corresponding cell.It checks in next layer of the GEM table of cell
Whether containing the attribute in subscribing to, if so, step 5 is then continued to execute, it otherwise will be in the S-node under subscription store to cell;
(6) check whether the subscription in S-node exceeds tolerance after storing.If exceeding, divided in S-node;Otherwise it inserts
Enter to terminate, building terminates.
(7) S-node is divided, and is counted to the attribute of subscription, is selected optimum attributes.Generate the GEM knot of this attribute
Structure extracts the subscription containing this attribute out from S-node, and storage is into newly-built GEM in corresponding cell.
6. the design of GEM-Tree event matches algorithm according to claim 1
The process that GEM-Tree carries out event matches is as follows:
(1) it according to the attribute of event and attribute number, is determined in attribute filter table and carries out matched grid;
(2) according to the attribute in event, the corresponding GEM of attribute is selected to be matched;
(3) on GEM structure (upper triangle), according to the value of attribute, cut-off rule delimited in upper triangle, cut-off rule pierced through
Cell screen out mismatch into event matches and subscribe to.
(4) traversal cell is connected to the list of next layer of GEM structure, if the attribute that contains of discovery event, repeat step 2,3,
4, until the list of the next layer of GEM structure of the cell at place is not present, or search less than respective attributes.
(5) all matched subscription are combined, obtains final matching result.
7. the contrast verification of GEM-Tree performance according to claim 1 has built information capacity up to million grades of reality
Platform is tested, from quantity is subscribed to, the matching degree of subscription, the size of attribute space, many factors such as attribute number of subscription are considered;Design
Several scenes carry out Experimental comparison and Performance Evaluation to matching algorithm;Matching is selected from tree structure and count matches algorithm
Algorithm AP-Tree, BE-Tree, OP-Index, REIN are reappeared, and carry out Experimental comparison with GEM-Tree algorithm.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811530822.5A CN109558410A (en) | 2018-12-14 | 2018-12-14 | Event matches algorithm based on multi-dimensional content in a kind of information distribution system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811530822.5A CN109558410A (en) | 2018-12-14 | 2018-12-14 | Event matches algorithm based on multi-dimensional content in a kind of information distribution system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109558410A true CN109558410A (en) | 2019-04-02 |
Family
ID=65870030
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811530822.5A Pending CN109558410A (en) | 2018-12-14 | 2018-12-14 | Event matches algorithm based on multi-dimensional content in a kind of information distribution system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109558410A (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103412883A (en) * | 2013-07-17 | 2013-11-27 | 中国人民解放军国防科学技术大学 | Semantic intelligent information publishing and subscribing method based on P2P technology |
CN105373633A (en) * | 2015-12-23 | 2016-03-02 | 江苏省现代企业信息化应用支撑软件工程技术研发中心 | Top-k subscription inquiring and matching method of position sensing subscription/publishing system |
US20160092567A1 (en) * | 2014-09-29 | 2016-03-31 | International Business Machines Corporation | Social media bot to representational state transfer (rest) proxy for data systems |
-
2018
- 2018-12-14 CN CN201811530822.5A patent/CN109558410A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103412883A (en) * | 2013-07-17 | 2013-11-27 | 中国人民解放军国防科学技术大学 | Semantic intelligent information publishing and subscribing method based on P2P technology |
US20160092567A1 (en) * | 2014-09-29 | 2016-03-31 | International Business Machines Corporation | Social media bot to representational state transfer (rest) proxy for data systems |
CN105373633A (en) * | 2015-12-23 | 2016-03-02 | 江苏省现代企业信息化应用支撑软件工程技术研发中心 | Top-k subscription inquiring and matching method of position sensing subscription/publishing system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104199986B (en) | Vector data space index method based on hbase and geohash | |
CN107423368B (en) | Spatio-temporal data indexing method in non-relational database | |
CN103678520B (en) | A kind of multi-dimensional interval query method and its system based on cloud computing | |
Zhang et al. | Scalable skyline computation using object-based space partitioning | |
CN106528773B (en) | Map computing system and method based on Spark platform supporting spatial data management | |
CN103678550B (en) | Mass data real-time query method based on dynamic index structure | |
CN103177094B (en) | Cleaning method of data of internet of things | |
CN107229751A (en) | A kind of concurrent incremental formula association rule mining method towards stream data | |
CN104504018B (en) | Based on dense tree and top-down big data real-time query optimization method | |
CN108009265B (en) | Spatial data indexing method in cloud computing environment | |
CN106209989A (en) | Spatial data concurrent computational system based on spark platform and method thereof | |
CN111639075B (en) | Non-relational database vector data management method based on flattened R tree | |
CN106951526B (en) | Entity set extension method and device | |
CN103377237A (en) | High dimensional data neighbor search method and fast approximate image search method | |
Hong et al. | Efficient R-tree based indexing scheme for server-centric cloud storage system | |
CN106874367A (en) | A kind of sampling distribution formula clustering method based on public sentiment platform | |
CN101430693B (en) | Spacing query method for triangular gridding curve model | |
CN112395288A (en) | R-tree index merging and updating method, device and medium based on Hilbert curve | |
CN111475511A (en) | Data storage method, data access method, data storage device, data access device and data access equipment based on tree structure | |
CN105138607B (en) | A kind of KNN querying methods based on combination grain distributed memory grid index | |
CN110069500A (en) | A kind of non-relational database dynamic hybrid index method | |
CN108681577A (en) | A kind of novel library structure data index method | |
CN109241236A (en) | Ocean geography Spatial Multi-Dimensional time-varying field data distribution formula tissue and inquiry processing method | |
CN109558410A (en) | Event matches algorithm based on multi-dimensional content in a kind of information distribution system | |
CN104794237B (en) | web information processing method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190402 |
|
WD01 | Invention patent application deemed withdrawn after publication |