CN103914563A - Pattern mining method for spatio-temporal track - Google Patents
Pattern mining method for spatio-temporal track Download PDFInfo
- Publication number
- CN103914563A CN103914563A CN201410157418.3A CN201410157418A CN103914563A CN 103914563 A CN103914563 A CN 103914563A CN 201410157418 A CN201410157418 A CN 201410157418A CN 103914563 A CN103914563 A CN 103914563A
- Authority
- CN
- China
- Prior art keywords
- space
- time
- spatio
- track
- frequent
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Fuzzy Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a pattern mining method for a spatio-temporal track. The pattern mining method for the spatio-temporal track includes: A, counting data according to the existing spatio-temporal points, and pre-processing the data according to the original information to generate the spatio-temporal track; B, carrying out pattern mining on the generated spatio-temporal track, wherein the mining method is characterized by redefining a spatio-temporal similarity measuring method and evolving the spatio-temporal similarity into spatio-temporal space for calculating; C, enabling a spatio-temporal similarity feature set to only comprise a time element and a space element; D, using a Prefi*Span method to obtain frequent items by means of the spatio-temporal similarity defined in the step B, wherein the frequent items comprise the frequent moving path and place of the user; E, using a convertible random agile storage mode-pseudo projection mode to store the gathered information; F, displaying the frequent items mined through the pattern mining method in a system platform. The pattern mining method for the spatio-temporal track uses the logic structure of the Prefi*Span method to guarantee the ordering of the spatio-temporal points in the track mode.
Description
Technical field
The present invention relates to a kind of data mining technology, more precisely a kind of in node motion situation the mode excavation method to space-time track.
Background introduction
The mode excavation of described space-time track, excavates a large amount of historical tracks of personal user, thereby finds out empty sequence when meeting threshold value and requiring (minimum frequent degree) frequent.Such as, expect a class user when advertiser and carry out advertisement delivery, where billboard is placed on to descried possibility maximum? and when learning the traffic conditions of current time on the A of section, does how the comprehensive section A user that above all users' driving information dopes the traffic of next moment section B and will enter to these section B that blocks up remind? in fact, also have various similar problems, but integrate be exactly: for someone, does is which bar maximum route that he/her walks? or, when having determined that starting point is A, when terminal is B, known his historical information, which bar route can most probable select? several hypothesis are the most realistic application that space-time trajectory model excavates above.
And spacetime correlation point analysis is the connectivity analysis laying particular emphasis between event, its realization based on obtained frequent time empty sequence.But existing conventional sequence mode excavation method all can not meet the demands.
For example, Apriori method is a kind of method of the most influential Mining Boolean Association Rules frequent item set.According to two fundamental propertys of collection, that is: 1) all nonvoid subsets of frequent item set must be frequent item sets; 2) superset of any non-frequent item set is also non-frequent item set.Like this, in mining process, just produce a large amount of candidates, and need to repeatedly scan the sequence library of respective stored data.
At current large data age, in the time that data volume constantly increases, just expose a huge shortcoming, because candidate is huge, need the data of storage many, the calculated amount of the method makes the memory space of database very large, and need to scan for a long time when the information in database of being stored in is inquired about, so will present exponential explosive increase.This is undoubtedly a task that can not realize for large batch of Mobile data.
And for example, FP growing method is that a kind of method of attempting by not producing candidate's frequent item set is called frequent pattern-growth, is called for short FP and increases.It is first by database compressing to frequent pattern tree (fp tree) (or FP tree) that provides frequent, but still a reservation collection related information.Then the database partition after compression is become to a set condition database (a kind of data for projection storehouse of special type), each associated one frequent, and excavate respectively each condition database.
Although this method efficiency has improved many, but the excavation for space-time trajectory model is still improper, because the mode internal obtaining in described method, the order of element does not have strict regulations, it has just reflected the sequencing on space, do not relate to time element, very high but space-time data excavates this some requirement.Such as someone appears at A ground the morning, and appear at C ground afternoon, appear at again B ground evening.In the storage of FP growing method, only has (ABC) this kind of storage mode, and we need be this storage mode of ACB, so this is the time empty sequence of a common-sense, also only just can be meaningful under time and spatial constraint, so in view of this this method is also inappropriate.
Find out from above-mentioned two kinds of methods, a kind of based on current data volume large, the novel track method for expressing that can comprise the time and space needs to be proposed, method based on pattern-growth not only can be excavated Frequent Sequential Patterns fast, and order that can also Assured Mode inside.Method based on pattern-growth is first to find out each frequent, then produces the set in data for projection storehouse, associated one frequent of each data for projection storehouse.Each database carries out independent excavation.Wherein best is exactly PrefixSpan method, i.e. prefix projection sequence pattern-growth.
In described method, the present invention intends adopting prefix pattern, associatedly with suffix pattern obtains frequent mode, thereby avoids producing a large amount of Candidate Sets, prevents from cannot implementing too greatly or need to costing a lot of money the time when scan database because of the data volume of required storage.Described method only detects the leading portion tract that is called prefix, by database projection in this prefix, excavates wherein frequent, then extends in prefix, continues to excavate, until excavate all frequent sequences.Be enhanced on spatiotemporal efficiency than Apriori method.Thereby be built into design of the present invention.
Summary of the invention
The object of the present invention is to provide a kind of mode excavation method of space-time track, the mode excavation method of a kind of space-time track of the present invention is the advantage of having utilized PrefixSpan method, redefine the measure of space-time similarity, continue to use the advantage of PrefixSpan method, do not produce candidate sequence, need the data volume of storage and make storage mode be associated with time element and the large characteristic of Spatial elements two thereby reduced.Which track is this method can judge in many tracks is the most similar, generates the most mobile route of users subsequently according to these tracks, is applied to subsequently in industrialization.
The method for digging of space-time trajectory model of the present invention, is characterized in that:
The mode excavation method of A, space-time track comprises the following steps:
1) according to carrying out data statistics existing event, specifically carry out data pre-service according to raw information, comprise data integration, data scrubbing, data transformation, thereby generate space-time track;
2) the space-time track of generation is carried out to mode excavation, the innovation of described method for digging has been to redefine the measure of space-time similarity, space-time similarity is developed into space-time distance and calculate;
3) feature set that redefines space-time similarity only comprises time and two, space element, thereby provides the method for space-time similarity;
4) method having thus described the invention is carried out the excavation of space-time trajectory model, continues to use the up-to-date described principle of PrefixSpan method, uses step 2) definition space-time similarity, obtain frequent, route, place that user frequently walks about;
5) by collect information storage time adopted a kind of disposable storage mode, i.e. pseudo-projecting method flexibly at random;
6) shown in system platform for frequent that finally the method with involved in the present invention is excavated;
B, described method have been utilized the advantage of PrefixSpan method, have redefined the measure of space-time similarity, continue to use the advantage of PrefixSpan algorithm, do not produce candidate sequence;
C, which track is described method can judge is the most similar, and comprehensive these tracks obtain mobile alignment the most frequently thereupon;
D, described frequent item i.e. frequent searched item out in this excavates.For example in supermarket, can excavate a certain commodity often purchased, in user's the route of walking about, the vehicle flowrate of which bar circuit is huge etc.So-called frequent item set, i.e. the set of all frequent, the frequent item set of buying article such as supermarket can be written as milk, bread or egg;
E, described the method at that time feature set of Kongxiang similarity only have two elements: time and space.
F, described method similarity are the one conversion of space-time distance, so the computing method of space-time similarity are:
SpatiTemporalDistance(Point1,Point2)=SpaceDistance(Point1,Poin?t2)*k+TimeDifference(Point1,Point2)*(1-k),0≤k≤1;
The core of G, described PrefixSpan method is to divide search volume with SDB, excavates respectively and contains the frequent sequence that length that these frequent sequences are prefix is K+1, until Result be sky;
H, described method can produce many data for projection storehouse, corresponding one of each subsequence of prefix frequently.Pseudo-projection can be recorded the reference position of projection suffix in the index (or identifier) of corresponding sequence and sequence, rather than sets up physics projection;
The physics projection of I, described sequence is recorded the identifier of sequence and the index point of projected position replaces;
J, described method are divided into three parts, Part I is the entrance of program, i.e. the carrying out of statement initialization sequence identifier, the main body of the method that Part II is recursive call, it is mainly to realize space-time trajectory model to excavate, and Part III is the realization in system platform;
The generation method of K, described track is:
1) existing point is carried out to data statistics, find frequent point;
2) position, adjacent point of time are merged, produce turnover time series; As, (A
i, starttime
i-endtime
i);
3) by this chopping rule, starting point and terminal are joined in track;
4) block track;
In L, described method, adopt the logical organization of PrefixSpan method, so can guarantee the order of event in trajectory model.While, can be utilized some fundamental propertys of frequent degree to may situation carrying out Quick tailoring, and utilize new self-defined space-time similarity measurement method, so be correct and efficient with Apriori character because of this algorithm.
Usually said space-time similarity is that two people or two tracks have all passed through A point (this has wherein not only comprised time element but also comprise Spatial elements) in some time points or certain time period determined, the evaluation of space-time similarity is similar apart from quantizing some to the feature set in the cluster analysis of data mining, and only in space-time similarity, feature set only has two elements: time and space.So in the present invention similarity namely space-time apart from the one conversion of SpatiTemporalDistance, so space-time distance B istance computing method are exactly SpatiTemporalDistance (Point1, Point2)=SpaceDistance (Point1, Point2) * k+TimeDifference (Point1, Point2) * (1-k), 0≤k≤1.In instantiation, k can adjust.
If sequence library is S, it is tuple <sequenceID, the set of sequence>.Wherein sequence is a sequence, and sequenceID is its numbering.And sequence can be defined as the ordered set <e of some events
1, e
2, e
3>, wherein each e
i(1≤i≤n) representative is an event in a sequence of events.Such as in the sequence library of shopping, S is all consumer records of user, and sequencei is the sequence of buying behavior in i article of record, e
jfor the j time consumer behavior in i article of record.
First carry out the relation of inclusion of defined nucleotide sequence, A is the subsequence of B, if there is integer 1≤j
1<j
2<j
m≤ n, makes
...,
.As A=<a, bc, d>, B=sequence=<sd, abc, bcn, f, sd>.Relation of inclusion refers in sequence A and meets above-mentioned subsequence definition, tuple <sequenceID, and sequence> has comprised sequence A, and in other words, sequence A once occurs in sequence library S.And the frequent degree of sequence A in sequence library S is the number of the tuple that comprises A in database.So, if certain sequence sequencePattern is a sequence pattern, be frequent sequence, must meet: the frequent degree of sequencePattern is greater than minimum frequent degree threshold value in S, the number of times that in S, sequencePattern occurs can be greater than certain given minimum value.
After having obtained initial raw data, data are carried out to pre-service, generate satisfactory space-time track sequence according to above-mentioned sequence pattern, generate the memory module of space-time track with reference to accompanying drawing 2.
A. track is carried out to segmentation, starting point and the terminal of definition track.If exceed certain period in the time in somewhere, this point is the terminal of current track, and if the time gap of the next point of this point and its while being no more than certain threshold values (Interval), it is also the starting point of next track;
B. redefining of track.Be different from the integrality of GPS, it lays particular emphasis on the participation of starting point and terminal;
C. the POI relevant with terminal according to starting point (Point of Interest point of interest), pays close attention to and excavates background knowledge;
Therefore T is as follows for definition track:
(A
1,starttime
1-endtime
1)(A
2,starttime
2-endtime
2)…(A
n,starttime
n-endtime
n).
Then the computing method that we continue to use Prefix method and above-mentioned space-time similarity are excavated the space-time track of above-mentioned formation, divide search volume with SDB, excavate respectively that to contain these frequent sequences be that prefix length is the frequent sequence of K+1, until Result be sky.
In this process, method of the present invention is characterised in that and in mining process, does not produce candidate sequence, but may generate many data for projection storehouse, corresponding one of each subsequence of prefix frequently.If projection must produce a new database, that thereupon can generate the database of a greater number.
As can be seen here, space-time track essence of the present invention is the sequence of an event.For space-time trajectory model, not only need to meet the requirement of General Sequences pattern, meanwhile, also to meet following characteristics:
1) the each point in sequence is with strict time character, as entry time, and time departure, duration;
2) be exactly an independently event e the each event in sequence
i;
3) e event in two sequences
i, e
jwhile coupling, traditional similarity measurement standard is inapplicable, so must define new distance-finding method;
4) different from general sequence pattern, space-time trajectory model must, according to background information, filter and screen obtained pattern in excavating;
Method of the present invention is in mining process, not produce candidate sequence, but may generate many data for projection storehouse, corresponding one of each subsequence of prefix frequently.If data for projection must physically produce, recurrence has built a large amount of data for projection storehouse and has just become the main expense of method of the present invention.
So adopted a kind of storage mode that is called pseudo-projection in method of the present invention, described pseudo-projection refers to that the physics projection of sequence is recorded the identifier of sequence and the index point of projected position replaces.When pseudo-projection can realize in internal memory, the expense of projection significantly reduces, if pseudo-projection access based on hard disk, just so ineffective.If initial trace database or data for projection storehouse are too large, can not be put in internal memory, should use physics projection, can be put in internal memory once data for projection storehouse, should use pseudo-projection.
Because adopted the logical organization of PrefixSpan method, so can guarantee the characteristic of time and space element combination in trajectory model, guarantee the order of event.Simultaneously because method of the present invention with Apriori character, can utilize some fundamental propertys of frequent degree to may situation carrying out Quick tailoring, and utilize new self-defined space-time similarity measurement method, so be correct and efficient, this is the practical part of this method just also, if adopt former classic method, first because memory data output is larger, be very difficult so implement.Secondly classic method can not guarantee the compossibility of time element and Spatial elements, usually attends to one thing and lose sight of another, and can not represent well place and route that user moves.So method of the present invention is not only practical but also efficient, but not pure computing method, though the content of the rule and methodology that the present invention comprises intellection, but comprise corresponding technical characterictic.
Accompanying drawing explanation
Fig. 1 is that space-time trajectory model excavates the process in whole data processing;
Fig. 2 is that space-time trajectory model Result is visual;
Fig. 3 space-time trajectory model excavates operational process.
Specific implementation method
Further illustrate substantive distinguishing features of the present invention and significant progressive below by accompanying drawing explanation and embodiment, but the present invention is only confined to by no means embodiment.
Fig. 1 has shown that novel space-time trajectory model excavates the process of deal with data, first the raw data of collecting (as mobile phone location data, gps signal data etc.) is carried out to data pre-service, this process comprises data integration, data scrubbing, data transformation, collected hash is converted into neat, unified database storage format, generates subsequently space-time track.Utilize novel space-time track method for digging to excavate frequent (place or the route that individual often walks about) of space-time track, and time row spacetime correlation analysis.Carry out context aware and prediction according to these information, and shown in system, by visual personnel location information (being shown on *** map).
According to Fig. 1 data mining process be:
1), read in sequence library SDB(SDB database-name for this reason) and minimum frequent degree threshold value (generally since 1);
2), default sequence length K=1 for the first time, from mapping database, find that length is the frequent sequence sets SDB of K, first Mining Frequent item is a database of frequent formation of a or b or c as shown in the figure, and frequent sequence is the sequence that in database, frequency is not less than threshold value;
3), excavate respectively and contain the frequent sequence that length that these frequent sequences are prefix is K+1, in the subdata base that is a in the prefix forming, add again one frequent, continue in subdata base, to search for the frequent item that prefix is ab, if Result is empty, stop;
4), sequence length K is increased to 1, by 3) L that finds gives SDB, then forwards 3 to);
5), record and export the frequent sequence that all excavations are arrived.
Fig. 2 has shown that user by name 91961 user is at the time number of times of several base station area position, and the route moving from him can be found out the scope that he walks about.
Fig. 3 has shown the process that space-time trajectory model excavates.Can be found out by this figure, the user of colony is at t
14 times, moment (or period) process A place, at t
24 times, moment (or period) process B place, at t
34 times, moment (or period) process C place.The space-time track that final raw data process generates is as shown in the block diagram of right side.
Method for digging of the present invention is divided into three parts, the entrance that Part I is program, the carrying out of statement initialization sequence identifier; Part II is the method main body of recursive call, is mainly to realize space-time trajectory model to excavate, and Part III is applied with real system platform.
Part I: GetFrequentSet is as follows for initialization homophony method:
1GetFrequentSet(ChosenNumber,setInterval)
2for i ← 1to sequenceSet.Count//initialization identifier array, is made as 03do initialIndex.Add (0)
4ProcessProjectedDB (intialIndex, setFrequency) // call analysis by start method
This step is that all marks are put in order to original position, and the original state of array is described, prepares to start to carry out track excavation.
Being described below of the space-time data method for digging of Part II based on pattern-growth:
ProcessProjectedDB(sequenceIndex,frequency,preStr)
The frequency that each class of 1for i ← 1to sequenceIndex.Count//first add up occurs event
2do if sequenceIndex[i] <sequenceSet[i] if .Count//identifier does not exceed the length of this sequence
3then for j ← 1sequenceIndex[i] to sequenceSet[i] .Count//statistics data for projection storehouse
4do?Count?the?number?of?each?space-time?point?as?a?Term.By?comparing?the?SpatiTemporalDistance(Point1,Point2)with?a?threshold?to?know?if?the?two?points?are?similar.
5for?each?term?in?termCount
6do?if?number(term)>frequency
When not existing, 7then if frequentSet.ContainsKey (preStr+term)=FALSE//pattern adds
8Then?frequentPatternSet.Add(preStr+term,number(term))
9for k ← 1to sequenceIndex.Count//more new identifier array
10do for1 ← sequenceIndex[k] to sequenceSet[k] the next index of .Count//find
11do?Find?the?index?of?current?term?in?sequenceSet[k]
12and?Get?the?new?sequenceIndex?to?newIndex
13ProcessProjectedDB (newIndex, frequency, preStr) // recursive call
First this step has added up the frequency of each class appearance event, if identifier event of adding up does not exceed the length of defined, identifier is projected in database and stored, the line number of going forward side by side Data preprocess, carries out data integration, data cleansing, data transformation.If event, identifier exceeded the length of defined, illustrate that this pattern does not exist, continue to find NeXt ID.Net result is stored in set of modes, and the form of each pattern is as follows:
(ci
1,time
1)(ci
2,time
2)(ci
a,time
a)……(ci
n,time
n):Number
Wherein ci
irepresent the label ID of base station of living in, time
ifor entering the time of this base station of user, Number represents the frequent degree (being which time in this excavates of this pattern occurs) of this pattern.
The application that Part III is combined with real system platform
Described method is applied to a real system platform by this part.In the time having a new data to arrive, can be joined in a database and be carried out unified management by data pre-service.Then trigger track by new data and generate processing module, obtain up-to-date mobile trajectory data.Data-mining module can be used as an independent processing logic, and it has comprised various excavations and has processed needed algorithm and controlling mechanism.Be the high efficiency of assurance system simultaneously, also some Results of this module deposited in database, to directly access while needing.
Claims (6)
1. a mode excavation method for space-time track, is characterized in that comprising the following steps:
A, according to carrying out data statistics existing event, carry out data pre-service according to raw information, thereby generate space-time track;
B, the space-time track of generation is carried out to mode excavation, described method for digging has been to redefine the measure of space-time similarity, space-time similarity is developed into space-time distance and calculate;
The feature set of C, space-time similarity only comprises time and two, space element;
D, continue to use PrefixSpan method, use the space-time similarity of step B definition, obtain frequent, route, place that user frequently walks about;
E, by collect information storage time adopted a kind of disposable storage mode flexibly at random, the i.e. storage mode of pseudo-projection;
F, finally by frequent shown in system platform of excavating by method involved in the present invention.
2. method according to claim 1, is characterized in that continuing to use PrefixSpan method, does not produce candidate sequence.
3. method according to claim 1, it is the most similar it is characterized in that judging which track, all crosses comprehensive these tracks thereupon and obtains mobile alignment the most frequently.
4. method according to claim 1, it is characterized in that in described method for digging frequent searched go out the set of frequent, be referred to as frequent item set.
5. method according to claim 1, it is characterized in that described similarity is the one conversion of space-time distance, so the computing method of space-time similarity are SpatiTemporalDistance (Point1, Point2)=SpaceDistance (Point1, Point2) * k+TimeDifference (Point1, Point2) * (1-k), 0≤k≤1.
6. method according to claim 1, is characterized in that described pseudo-projection refers to that the physics projection of sequence is recorded the identifier of sequence and the index point of projected position replaces.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410157418.3A CN103914563A (en) | 2014-04-18 | 2014-04-18 | Pattern mining method for spatio-temporal track |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410157418.3A CN103914563A (en) | 2014-04-18 | 2014-04-18 | Pattern mining method for spatio-temporal track |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103914563A true CN103914563A (en) | 2014-07-09 |
Family
ID=51040243
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410157418.3A Pending CN103914563A (en) | 2014-04-18 | 2014-04-18 | Pattern mining method for spatio-temporal track |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103914563A (en) |
Cited By (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104331466A (en) * | 2014-10-31 | 2015-02-04 | 南京邮电大学 | Space-time proximity search-based mobile trace sequence mode quick mining method |
CN104754509A (en) * | 2014-10-31 | 2015-07-01 | 南京邮电大学 | LBS (Location Based Service) privacy protecting method based on position semantics K-anonymity |
CN105117424A (en) * | 2015-07-31 | 2015-12-02 | 中国科学院软件研究所 | Dwell-time-based moving object semantic behavior pattern mining method |
CN105224688A (en) * | 2015-10-30 | 2016-01-06 | 宁波大学 | A kind of Apriori method being applicable to excavation mobile node motion track correlation rule |
CN105224687A (en) * | 2015-10-30 | 2016-01-06 | 宁波大学 | A kind of low consumed Apriori method |
CN105592415A (en) * | 2015-11-04 | 2016-05-18 | 宁波大学 | Low-consumed FMIPv6 switching method on the basis of mobile predication |
WO2016127507A1 (en) * | 2015-02-12 | 2016-08-18 | 中兴通讯股份有限公司 | Method and apparatus for analyzing track of user |
CN106485921A (en) * | 2016-12-07 | 2017-03-08 | 北京航空航天大学 | A kind of region highway network wagon flow state analysis method based on social data |
CN106649656A (en) * | 2016-12-13 | 2017-05-10 | 中国科学院软件研究所 | Spatial-temporal trajectory big data storage method for database |
CN106778876A (en) * | 2016-12-21 | 2017-05-31 | 广州杰赛科技股份有限公司 | User classification method and system based on mobile subscriber track similitude |
CN106844736A (en) * | 2017-02-13 | 2017-06-13 | 北方工业大学 | Time-space co-occurrence mode mining method based on time-space network |
CN107016495A (en) * | 2017-03-21 | 2017-08-04 | 乐蜜科技有限公司 | Determination method, device and the terminal device of urban area correlation |
CN107194434A (en) * | 2017-06-16 | 2017-09-22 | 中国矿业大学 | A kind of mobile object similarity calculating method and system based on space-time data |
CN107680010A (en) * | 2017-09-29 | 2018-02-09 | 桂林电子科技大学 | A kind of scenic spot route recommendation method and its system based on visit behavior |
CN107679558A (en) * | 2017-09-19 | 2018-02-09 | 电子科技大学 | A kind of user trajectory method for measuring similarity based on metric learning |
WO2018219057A1 (en) * | 2017-06-01 | 2018-12-06 | 华为技术有限公司 | Site selection method and device |
CN109325635A (en) * | 2018-10-25 | 2019-02-12 | 电子科技大学中山学院 | Position prediction method based on automatic completion |
US10417648B2 (en) | 2015-11-09 | 2019-09-17 | Industrial Technology Research Institute | System and computer readable medium for finding crowd movements |
CN110309434A (en) * | 2018-10-10 | 2019-10-08 | 腾讯大地通途(北京)科技有限公司 | A kind of track data processing method, device and relevant device |
CN112182410A (en) * | 2020-11-05 | 2021-01-05 | 西南交通大学 | User travel mode mining method based on space-time trajectory knowledge graph |
CN114530018A (en) * | 2022-04-24 | 2022-05-24 | 浙江华眼视觉科技有限公司 | Voice prompt method and device based on pickup trajectory analysis |
CN115795115A (en) * | 2023-02-11 | 2023-03-14 | 云南师范大学 | Multi-track set similarity searching method based on graph storage |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102646070A (en) * | 2012-02-29 | 2012-08-22 | 武汉大学 | Space-time trajectory data storage method based on area |
CN102722541A (en) * | 2012-05-23 | 2012-10-10 | 中国科学院计算技术研究所 | Method and system for calculating space-time locus similarity |
CN102855638A (en) * | 2012-08-13 | 2013-01-02 | 苏州大学 | Detection method for abnormal behavior of vehicle based on spectrum clustering |
CN103593361A (en) * | 2012-08-14 | 2014-02-19 | 中国科学院沈阳自动化研究所 | Movement space-time trajectory analysis method in sense network environment |
-
2014
- 2014-04-18 CN CN201410157418.3A patent/CN103914563A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102646070A (en) * | 2012-02-29 | 2012-08-22 | 武汉大学 | Space-time trajectory data storage method based on area |
CN102722541A (en) * | 2012-05-23 | 2012-10-10 | 中国科学院计算技术研究所 | Method and system for calculating space-time locus similarity |
CN102855638A (en) * | 2012-08-13 | 2013-01-02 | 苏州大学 | Detection method for abnormal behavior of vehicle based on spectrum clustering |
CN103593361A (en) * | 2012-08-14 | 2014-02-19 | 中国科学院沈阳自动化研究所 | Movement space-time trajectory analysis method in sense network environment |
Non-Patent Citations (1)
Title |
---|
龚玺等: "时空轨迹聚类方法研究进展", 《地理科学进展》 * |
Cited By (36)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104331466A (en) * | 2014-10-31 | 2015-02-04 | 南京邮电大学 | Space-time proximity search-based mobile trace sequence mode quick mining method |
CN104754509A (en) * | 2014-10-31 | 2015-07-01 | 南京邮电大学 | LBS (Location Based Service) privacy protecting method based on position semantics K-anonymity |
CN104331466B (en) * | 2014-10-31 | 2018-01-19 | 南京邮电大学 | Motion track sequence pattern Fast Mining Algorithm based on space-time proximity search |
CN104754509B (en) * | 2014-10-31 | 2018-04-06 | 南京邮电大学 | Based on the anonymous LBS method for secret protection of position semanteme K |
WO2016127507A1 (en) * | 2015-02-12 | 2016-08-18 | 中兴通讯股份有限公司 | Method and apparatus for analyzing track of user |
CN105117424A (en) * | 2015-07-31 | 2015-12-02 | 中国科学院软件研究所 | Dwell-time-based moving object semantic behavior pattern mining method |
CN105117424B (en) * | 2015-07-31 | 2019-08-02 | 中国科学院软件研究所 | A kind of mobile object semanteme behavior patterns mining method based on the residence time |
CN105224687A (en) * | 2015-10-30 | 2016-01-06 | 宁波大学 | A kind of low consumed Apriori method |
CN105224688A (en) * | 2015-10-30 | 2016-01-06 | 宁波大学 | A kind of Apriori method being applicable to excavation mobile node motion track correlation rule |
CN105224688B (en) * | 2015-10-30 | 2018-10-19 | 宁波大学 | A kind of Apriori methods being suitable for excavating mobile node motion track correlation rule |
CN105224687B (en) * | 2015-10-30 | 2018-10-19 | 宁波大学 | A kind of low consumed Apriori methods |
CN105592415A (en) * | 2015-11-04 | 2016-05-18 | 宁波大学 | Low-consumed FMIPv6 switching method on the basis of mobile predication |
CN105592415B (en) * | 2015-11-04 | 2019-06-07 | 宁波大学 | Low consumption FMIPv6 switching method based on moving projection |
US10417648B2 (en) | 2015-11-09 | 2019-09-17 | Industrial Technology Research Institute | System and computer readable medium for finding crowd movements |
CN106485921A (en) * | 2016-12-07 | 2017-03-08 | 北京航空航天大学 | A kind of region highway network wagon flow state analysis method based on social data |
CN106485921B (en) * | 2016-12-07 | 2019-03-29 | 北京航空航天大学 | A kind of region highway network wagon flow state analysis method based on social data |
CN106649656B (en) * | 2016-12-13 | 2020-03-17 | 中国科学院软件研究所 | Database-oriented space-time trajectory big data storage method |
CN106649656A (en) * | 2016-12-13 | 2017-05-10 | 中国科学院软件研究所 | Spatial-temporal trajectory big data storage method for database |
CN106778876A (en) * | 2016-12-21 | 2017-05-31 | 广州杰赛科技股份有限公司 | User classification method and system based on mobile subscriber track similitude |
CN106844736B (en) * | 2017-02-13 | 2021-07-16 | 北方工业大学 | Time-space co-occurrence mode mining method based on time-space network |
CN106844736A (en) * | 2017-02-13 | 2017-06-13 | 北方工业大学 | Time-space co-occurrence mode mining method based on time-space network |
CN107016495A (en) * | 2017-03-21 | 2017-08-04 | 乐蜜科技有限公司 | Determination method, device and the terminal device of urban area correlation |
WO2018219057A1 (en) * | 2017-06-01 | 2018-12-06 | 华为技术有限公司 | Site selection method and device |
CN107194434A (en) * | 2017-06-16 | 2017-09-22 | 中国矿业大学 | A kind of mobile object similarity calculating method and system based on space-time data |
CN107679558B (en) * | 2017-09-19 | 2019-09-24 | 电子科技大学 | A kind of user trajectory method for measuring similarity based on metric learning |
CN107679558A (en) * | 2017-09-19 | 2018-02-09 | 电子科技大学 | A kind of user trajectory method for measuring similarity based on metric learning |
CN107680010B (en) * | 2017-09-29 | 2020-12-18 | 桂林电子科技大学 | Scenic spot route recommendation method and system based on touring behavior |
CN107680010A (en) * | 2017-09-29 | 2018-02-09 | 桂林电子科技大学 | A kind of scenic spot route recommendation method and its system based on visit behavior |
CN110309434A (en) * | 2018-10-10 | 2019-10-08 | 腾讯大地通途(北京)科技有限公司 | A kind of track data processing method, device and relevant device |
CN110309434B (en) * | 2018-10-10 | 2023-10-13 | 腾讯大地通途(北京)科技有限公司 | Track data processing method and device and related equipment |
CN109325635A (en) * | 2018-10-25 | 2019-02-12 | 电子科技大学中山学院 | Position prediction method based on automatic completion |
CN109325635B (en) * | 2018-10-25 | 2022-02-15 | 电子科技大学中山学院 | Position prediction method based on automatic completion |
CN112182410A (en) * | 2020-11-05 | 2021-01-05 | 西南交通大学 | User travel mode mining method based on space-time trajectory knowledge graph |
CN112182410B (en) * | 2020-11-05 | 2023-04-21 | 西南交通大学 | User travel mode mining method based on space-time track knowledge graph |
CN114530018A (en) * | 2022-04-24 | 2022-05-24 | 浙江华眼视觉科技有限公司 | Voice prompt method and device based on pickup trajectory analysis |
CN115795115A (en) * | 2023-02-11 | 2023-03-14 | 云南师范大学 | Multi-track set similarity searching method based on graph storage |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103914563A (en) | Pattern mining method for spatio-temporal track | |
US10558898B2 (en) | System and method for generating a map from activity data | |
Milan et al. | MOT16: A benchmark for multi-object tracking | |
Shen et al. | Multiobject tracking by submodular optimization | |
Yan et al. | Semantic trajectories: Mobility data computation and annotation | |
Yoon et al. | Cityride: a predictive bike sharing journey advisor | |
Yan et al. | SeMiTri: a framework for semantic annotation of heterogeneous trajectories | |
Zheng et al. | Learning location correlation from gps trajectories | |
Ceikute et al. | Routing service quality--local driver behavior versus routing services | |
KR101369261B1 (en) | Method for providing real-time route inference of users using smart terminal and system there of | |
US10754040B2 (en) | System and method for graph encoding of physical activity data | |
Banerjee et al. | Inferring uncertain trajectories from partial observations | |
CN105512727B (en) | A kind of personal path Forecasting Methodology based on Markov | |
CN109684373B (en) | Key relation person discovery method based on travel and call ticket data analysis | |
CN112668525B (en) | People flow counting method and device, electronic equipment and storage medium | |
CN108133619A (en) | Method, apparatus, storage medium and the terminal device of Parking prediction | |
Krueger et al. | Semantic enrichment of movement behavior with foursquare–a visual analytics approach | |
CN109190656B (en) | Indoor semantic track marking and complementing method under low-sampling positioning environment | |
KR102073208B1 (en) | stadium visitor big-data analysis system | |
CN111523577A (en) | Mass trajectory similarity calculation method based on improved LCSS algorithm | |
CN107615272A (en) | System and method for predicting crowd's attribute | |
CN110275911A (en) | Private car trip hotspot path method for digging based on Frequent Sequential Patterns | |
CN107844602A (en) | A kind of Forecasting Methodology based on time-space attribute correlation rule | |
CN109558961B (en) | Method and system for determining position information, storage medium, processor and device | |
CN109785656B (en) | Traffic planning and navigation method based on local approximate rank |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20140709 |
|
WD01 | Invention patent application deemed withdrawn after publication |