CN108182446A - A kind of driver's permanent residence Forecasting Methodology and device based on clustering algorithm - Google Patents
A kind of driver's permanent residence Forecasting Methodology and device based on clustering algorithm Download PDFInfo
- Publication number
- CN108182446A CN108182446A CN201711325708.4A CN201711325708A CN108182446A CN 108182446 A CN108182446 A CN 108182446A CN 201711325708 A CN201711325708 A CN 201711325708A CN 108182446 A CN108182446 A CN 108182446A
- Authority
- CN
- China
- Prior art keywords
- cluster
- submodule
- current
- point
- module
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Abstract
The invention discloses a kind of driver's permanent residence Forecasting Methodologies and device based on clustering algorithm, belong to intelligent transportation big data field.The method includes：The position data for collecting driver forms data acquisition system；DBSCAN algorithms are optimized, and cluster operation is carried out to the data acquisition system of collection using the DBSCAN algorithms of optimization, obtain the first cluster result；K means algorithms are optimized, based on the first cluster result, cluster operation carries out the data acquisition system of collection using the K means algorithms of optimization and obtains the second cluster result；According to the second cluster result, the permanent residence of driver is predicted.In the present invention, by the way that DBSCAN and K means algorithms are optimized and are combined, not only increase the computational efficiency in cluster process, and it ensure that the of overall importance and high accuracy of cluster result, so as to provide correct guidance instruction for the reference person of data, promote the development of industry and relevant industries.
Description
Technical field
The present invention relates to intelligent transportation big data field more particularly to a kind of driver's permanent residence predictions based on clustering algorithm
Method and device.
Background technology
Today's society is the society of a high speed development, and science and technology is flourishing, information flow, and the exchange between people is closer and closer
It cuts, life is also more and more convenient.Product of the big data as this cyberage, the more and more extensive life for coming into people
It is living, and applied to all trades and professions.The transportation that particularly people's trip and cargo transport are be unable to do without, it will usually based on department
Machine or a large amount of geographic position data of vehicle carry out space clustering, by the region of appearance intensive in cluster, are divided into permanent residence,
So as to provide foundation for the propulsion of the marketing program of vehicle scheduling, vehicle management and relevant industries and siteselecting planning etc..Its
In, common clustering algorithm is DBSCAN algorithms.It, can be due to when data volume is very big however, existing DBSCAN clustering algorithms
The Density inhomogeneity of space clustering, cluster pitch difference differ greatly, cause parameter Eps (radius of neighbourhood, it is true by core of current point
Determine length cited during density area range) and minPts (density threshold, using current point as core, using radius of neighbourhood Eps as
In the density area that length determines, meet the minimum element number of specified requirement) difficulty is chosen, so as to reduce clustering result quality,
Making cluster result and actual conditions, there are deviations, and correctly foundation is instructed so as to be provided to the reference person of data.
Invention content
I solves the deficiencies in the prior art, the present invention provide a kind of driver's permanent residence Forecasting Methodology based on clustering algorithm and
Device.
On the one hand, the present invention provides a kind of driver's permanent residence Forecasting Methodology based on clustering algorithm, including：
Step S1：The position data for collecting driver forms data acquisition system；
Step S2：DBSCAN algorithms are optimized, and the data acquisition system is carried out using the DBSCAN algorithms of optimization
Operation is clustered, obtains the first cluster result；
Step S3：K-means algorithms are optimized, based on first cluster result, are calculated using the K-means of optimization
Method carries out the data acquisition system cluster operation and obtains the second cluster result；
Step S4：According to second cluster result, the permanent residence of the driver is predicted.
Optionally, the step S1, specially：Collect the coordinate that client used in driver reports in preset time period
Point forms coordinate point set.
Optionally, it is described that DBSCAN algorithms are optimized in the step S2, it specifically includes：
Optimization the radius of neighbourhood meaning be：A certain current point adds in the basis for estimation of a certain cluster set；
The meaning of Optimal Density threshold value is：A certain cluster set synthesizes the basis for estimation to cluster；
Accordingly, in the step S2, the DBSCAN algorithms using optimization carry out cluster fortune to the data acquisition system
It calculates, obtains the first cluster result, specifically include：
Step A1：A pending coordinate points are randomly choosed in the coordinate point set as current point, are judged current
The quantity of set is clustered, the quantity of such as described current cluster set is zero, then performs step A2；Such as the current cluster set
Quantity is 1, then performs step A3；Quantity such as the current cluster set is more than 1, then performs step A4；
Step A2：Using the current point as the central point of first cluster set, and described first cluster is gathered
As current cluster set, return to step A1；
Step A3：Judge the distance between the current point and the central point of the current cluster set whether no more than institute
The radius of neighbourhood is stated, is then to gather current point cluster to the current cluster, and update in the current cluster set
Heart point performs step A5；Otherwise, cluster set is created, using the current point as the central point of newly-built cluster set, and will be new
The cluster set cooperation built is gathered for current cluster, performs step A5；
Step A4：The distance between the current point and the central point of each current cluster set are calculated, therefrom selection is minimum
Distance judges that the minimum range is then by current point cluster to the most narrow spacing whether no more than the radius of neighbourhood
From in corresponding current cluster set, the central point of the corresponding current cluster set of the minimum range is updated, performs step A5；
Otherwise, cluster set is created, using the central point that the current point is gathered as newly-built cluster, and the cluster set cooperation that will be created
Currently to cluster set, step A5 is performed；
Step A5：Judge in the coordinate set whether to also have untreated coordinate points, be then return to step A1；Otherwise, it holds
Row step A6：
Step A6：The density of each cluster set of generation is compared, and by density not with the density threshold successively
Cluster set cooperation less than the density threshold clusters for first, using corresponding central point as the first cluster centre, exports institute
It states first to cluster and corresponding first cluster centre, obtains the first cluster result.
Optionally, it is described that K-means algorithms are optimized in the step S3, specially：Radius parameter is set；
Accordingly, it is described based on first cluster result in the step S3, using the K-means algorithms pair of optimization
The data acquisition system carries out cluster operation and obtains the second cluster result, specifically includes：
Step B1：The second cluster centre is initialized as each first cluster centre in first cluster result, and by institute
The second cluster centre is stated as current goal cluster centre, using each first where the current goal cluster centre cluster as
Currently cluster；
Step B2：Traverse the coordinate points not traversed in the coordinate point set, calculate the coordinate points that traverse with it is each current
Whether the distance between target cluster centre, therefrom selects minimum range, judge the minimum range no more than the radius set
Parameter is to perform step B3；Otherwise, the coordinate points of traversal are given up, performs step B4；
Step B3：By working as where the coordinate points cluster to the corresponding current goal cluster centre of the minimum range of traversal
Before cluster after as currently clustering, update the cluster centre currently to cluster, and using updated cluster centre as current mesh
Cluster centre is marked, performs step B4；
Step B4：Judge it is then return to step B2 with the presence or absence of the coordinate points that do not traverse in the coordinate point set；Otherwise
Each current cluster is clustered as second, using each current goal cluster centre as the second cluster centre, output described second gathers
Cluster and corresponding each second cluster centre, obtain the second cluster result.
Optionally, the step S4, specially：Each second cluster centre that will contain in second cluster result is made
The permanent residence of the driver for prediction.
On the other hand, the present invention provides a kind of driver's permanent residence prediction meanss based on clustering algorithm, including：
Collection module forms data acquisition system for collecting the position data of driver；
First optimization module, for being optimized to DBSCAN algorithms；
First cluster module, for using first optimization module optimization DBSCAN algorithms to the data acquisition system into
Row cluster operation, obtains the first cluster result；
Second optimization module, for being optimized to K-means algorithms；
Second cluster module, for the first cluster result obtained based on first cluster module, using the second optimization
The K-means algorithms of module optimization carry out cluster operation to the data acquisition system that the collection module is collected, and obtain the first cluster knot
Fruit；
Prediction module for the second cluster result obtained according to second cluster module, predicts that the driver's is normal
Guard station.
Optionally, the collection module, is specifically used for：Collect visitor used in the corresponding driver of vehicle in preset time period
The coordinate points that family end reports form coordinate point set.
Optionally, first optimization module, is specifically used for：Optimization the radius of neighbourhood meaning be：A certain current point adds in
The basis for estimation of a certain cluster set；The meaning of Optimal Density threshold value is：A certain cluster set synthesizes the basis for estimation to cluster；
Accordingly, first cluster module, specifically includes：Select submodule, the first judging submodule, as submodule
Block, second judgment submodule, the first cluster submodule, the first newly-built submodule, the first computational submodule, third judge submodule
Block, the second newly-built submodule, the 4th judging submodule, compares submodule and the first output sub-module at the second cluster submodule；
The selection submodule, it is pending for random selection one in the coordinate point set collected in the collection module
Coordinate points are as current point；
First judging submodule, for judging currently to cluster the quantity of set；
It is described to be used as submodule, it is zero for working as the quantity that first judging submodule is judged currently to cluster set
When, using the current point that submodule is selected to select as the central point of first cluster set, and described first is clustered
Set is gathered as current cluster, triggers the selection submodule；
The second judgment submodule, for working as the number that first judging submodule judges the current cluster set
It measures when being 1, judges that the distance between central point that the current point of the selection submodule selection is gathered with the current cluster is
The no radius of neighbourhood no more than first optimization module optimization；
The first cluster submodule judges the selection submodule selection for working as the second judgment submodule
The distance between current point and the central point of the current cluster set are no more than the neighborhood half of first optimization module optimization
During diameter, the current point cluster of the selection submodule selection is gathered, and update the current cluster set to the current cluster
The central point of conjunction, triggering the 4th selection submodule；
The first newly-built submodule judges the selection submodule selection for working as the second judgment submodule
The distance between current point and the central point of the current cluster set are more than the radius of neighbourhood of first optimization module optimization
When, create cluster set；
The first update submodule is additionally operable to using the current point that submodule is selected to select as newly-built cluster set
Central point, and newly-built cluster set cooperation is gathered for current cluster, triggering the 4th selection submodule；
First computational submodule, for working as the number that first judging submodule judges the current cluster set
When amount is more than 1, the distance between described central point that the current point that submodule selects is selected to gather with each current cluster of calculating, from
Middle selection minimum range；
The third judging submodule, for judging whether the minimum range that first computational submodule obtains is not more than
The radius of neighbourhood of the first optimization module optimization；
The second cluster submodule, judges that first computational submodule obtains for working as the third judging submodule
When the minimum range arrived is no more than the radius of neighbourhood that first optimization module optimizes, by the current of the selection submodule selection
In point cluster to the corresponding current cluster set of the minimum range, the corresponding current cluster set of the minimum range is updated
Central point triggers the 4th judging submodule；
The second newly-built submodule judges that first computational submodule obtains for working as the third judging submodule
When the minimum range arrived is more than the radius of neighbourhood of first optimization module optimization, cluster set is created；
The second cluster submodule, is additionally operable to create the current point that submodule is selected to select as described second
The central point of cluster set that submodule creates, and newly-built cluster set cooperation is gathered for current cluster, triggering the described 4th
Judging submodule；
4th judging submodule, it is whether also untreated in the coordinate set of the collection module collection for judging
Coordinate points；
The selection submodule is additionally operable to judge the coordinate of the collection module collection when the 4th judging submodule
When also having untreated coordinate points in set, a pending seat is randomly choosed in the coordinate point set collected in the collection module
Punctuate is as current point；
The comparison submodule, for working as the coordinate set that the 4th judging submodule judges the collection module collection
When there is no untreated coordinate points in conjunction, successively by the density of each cluster set and the density threshold of first optimization module optimization
Be compared, and cluster set cooperation of the density not less than the density threshold clustered for first, using corresponding central point as
First cluster centre；
First output sub-module gathers for exporting the comparison submodule obtains first and clustering with corresponding first
Class center obtains the first cluster result.
Optionally, second optimization module, is specifically used for：Radius parameter is set；
Accordingly, second cluster module, specifically includes：Initialization submodule, traversal submodule, second calculate submodule
Block, third cluster submodule, gives up submodule, the 6th judging submodule and the second output sub-module at the 5th judging submodule；
The initialization submodule, for initializing the second cluster centre first for first output sub-module output
Each first cluster centre in cluster result, and using second cluster centre as current goal cluster centre, described will work as
Each first where preceding target cluster centre clusters as currently clustering；
The traversal submodule, for traversing the coordinate points not traversed in the coordinate point set of the collection module collection；
Second computational submodule is gathered for calculating the coordinate points that the traversal submodule traverses with each current goal
The distance between class center, therefrom selects minimum range；
5th judging submodule, for judging whether the minimum range that second computational submodule calculates is not more than
The radius parameter of the second optimization module setting；
The third clusters submodule, judges based on second computational submodule by working as the 5th judging submodule
When the minimum range of calculation is no more than the radius parameter that second optimization module is set, by the coordinate of the traversal submodule traversal
Point cluster cluster to current where the corresponding current goal cluster centre of the minimum range after as currently clustering, update and be somebody's turn to do
The cluster centre currently to cluster, and using updated cluster centre as current goal cluster centre, triggering the described 6th judges
Submodule；
It is described to give up submodule, judge what second computational submodule calculated for working as the 5th judging submodule
When minimum range is more than the radius parameter of second optimization module setting, the coordinate points of the traversal submodule traversal are given up
It abandons, triggers the 6th judging submodule；
6th judging submodule, for judging in coordinate point set that the collection module collects with the presence or absence of not time
The coordinate points gone through；
The traversal submodule, for working as the coordinate points that the 6th judging submodule judges the collection module collection
When there are the coordinate points not traversed in set, the coordinate points not traversed in the coordinate point set that the collection module is collected are traversed；
Second output sub-module, for working as the seat that the 6th judging submodule judges the collection module collection
There is no during the coordinate points not traversed in punctuate set, each current cluster is clustered as second, during each current goal is clustered
For the heart as the second cluster centre, output described second clusters and corresponding each second cluster centre, obtains the second cluster result.
Optionally, the prediction module, is specifically used for：Contain in the second cluster result that second cluster module is obtained
Each second cluster centre having, the permanent residence of the driver as prediction.
The advantage of the invention is that：
In the present invention, by the way that DBSCAN clustering algorithms and K-means algorithms are optimized and combined, i.e., based on optimization
DBSCAN algorithms carry out the position data of collection first time cluster, and the K- by the result of first time cluster as an optimization
The input data of means algorithms carries out second and clusters, not only increases computational efficiency, and ensure that the complete of cluster result
Office's property and high accuracy, and then correct guidance instruction can be provided for the reference person of data, promote industry and associated row
The development of industry.
Description of the drawings
By reading the detailed description of hereafter preferred embodiment, it is various other the advantages of and benefit it is common for this field
Technical staff will become clear.Attached drawing is only used for showing the purpose of preferred embodiment, and is not considered as to the present invention
Limitation.And throughout the drawings, the same reference numbers will be used to refer to the same parts.In the accompanying drawings：
Attached drawing 1 is a kind of driver's permanent residence Forecasting Methodology flow chart based on clustering algorithm provided by the invention；
The schematic diagram that attached drawing 2 is clustered for the DBSCAN algorithms provided by the invention using optimization；
The schematic diagram that attached drawing 3 is clustered for the K-means algorithms provided by the invention using optimization；
Attached drawing 4 is a kind of driver's permanent residence prediction meanss module composition frame chart based on clustering algorithm provided by the invention.
Specific embodiment
The illustrative embodiments of the disclosure are more fully described below with reference to accompanying drawings.Although this public affairs is shown in attached drawing
The illustrative embodiments opened, it being understood, however, that may be realized in various forms the disclosure without the reality that should be illustrated here
The mode of applying is limited.It is to be able to be best understood from the disclosure, and can be by this public affairs on the contrary, providing these embodiments
The range opened completely is communicated to those skilled in the art.
Embodiment one
According to the embodiment of the present invention, a kind of driver's permanent residence Forecasting Methodology based on clustering algorithm is provided, such as Fig. 1 institutes
Show, including：
Step 101：The position data for collecting driver forms data acquisition system；
Preferably, in the present embodiment, driver by client (for example, vehicle management APP) every prefixed time interval
Report the information such as coordinate points, the travel speed of oneself；Wherein, prefixed time interval, can sets itself according to demand, for example,
It is set as 60 seconds.
Accordingly, step 101, specially：The coordinate points that client used in driver reports in preset time period are collected,
Form coordinate point set；Wherein, preset time period, can sets itself according to demand, for example, being set as one week.
Step 102：DBSCAN algorithms are optimized, and data acquisition system is clustered using the DBSCAN algorithms of optimization
Operation obtains the first cluster result；
According to the embodiment of the present invention, DBSCAN algorithms are optimized, specifically included：
Optimization the radius of neighbourhood meaning be：A certain current point adds in the basis for estimation of a certain cluster set；
The meaning of Optimal Density threshold value is：A certain cluster set synthesizes the basis for estimation to cluster；
Accordingly, in step 102, cluster operation is carried out to data acquisition system using the DBSCAN algorithms of optimization, obtains first
Cluster result specifically includes：
Step A1：A pending coordinate points are randomly choosed in the coordinate point set of collection as current point, judge to work as
The quantity of preceding cluster set, the quantity of such as current cluster set is zero, then performs step A2；As the current quantity for clustering set is
1, then perform step A3；Quantity such as current cluster set is more than 1, then performs step A4；
Step A2：Using current point as the central point of first cluster set, and it is current by first cluster set cooperation
Cluster set, return to step A1；
In the present embodiment, when the quantity of current cluster set is zero, show that the current point selected is waited to locate for first
Shown in (a) in reason coordinate points, the then central point directly gathered as first cluster, schematic diagram such as Fig. 2.
Step A3：Judge whether the distance between current point and the central point of current cluster set are not more than the radius of neighbourhood,
It is then to gather current point cluster to current cluster, and update the central point of current cluster set, execution step A5；Otherwise, newly
Cluster set is built, using current point as the central point of newly-built cluster set, and is current cluster set by newly-built cluster set cooperation
It closes, performs step A5；
Wherein, the central point of the current cluster set of update, specially：Delineation can include each coordinate in current cluster set
The minimum circle of point, and using the center of circle as the central point of current cluster set.
In the present embodiment, illustrated so that the current point selected is second pending coordinate points as an example, when judge work as
It, ought when whether the distance between preceding point and central point of current cluster set (first cluster is gathered) are not more than the radius of neighbourhood
Preceding point cluster is gathered to current cluster, and updates the central point of current cluster set, shown in (b) in schematic diagram such as Fig. 2.
Step A4：The distance between current point and the central point of each current cluster set are calculated, therefrom selects minimum range,
Judge that minimum range is then by current point cluster to the corresponding current cluster of minimum range whether no more than the radius of neighbourhood optimized
In set, the central point of the corresponding current cluster set of update minimum range performs step A5；Otherwise, cluster set is created, it will
Central point of the current point as newly-built cluster set, and newly-built cluster set cooperation is gathered for current cluster, perform step
A5；
Preferably, the distance between current point and the central point of each current cluster set are calculated, specially：According to current point
With it is each it is current cluster set central point coordinate, calculate current point and it is each it is current cluster set central point between Euclidean away from
From.
In the present invention, the quantity by current cluster set illustrates for being 2, optimizes when judging that minimum range is more than
The radius of neighbourhood, cluster set is created, using current point as the schematic diagram of the central point of newly-built cluster set as in Fig. 2
(c) shown in.
Step A5：Judge in coordinate set whether to also have untreated coordinate points, be then return to step A1；Otherwise, step is performed
Rapid A6：
Step A6：The density of each cluster set of generation is compared with the density threshold of optimization successively, and by density
Cluster set cooperation not less than density threshold clusters for first, using corresponding central point as the first cluster centre, output first
It clusters and corresponding first cluster centre, obtains the first cluster result.
In the present invention, by being optimized to the greatest extent to DBSCAN algorithms, containing for the parameter radius of neighbourhood and density threshold is changed
Justice and purposes, so as to substantially increase the computational efficiency in cluster process.
Step 103：K-means algorithms are optimized, based on the first cluster result, using the K-means algorithms of optimization
Cluster operation is carried out to data acquisition system, obtains the second cluster result；
According to the embodiment of the present invention, K-means algorithms are optimized, specially：Radius parameter is set；
Accordingly, in step 103, based on the first cluster result, data acquisition system is carried out using the K-means algorithms of optimization
Cluster operation obtains the second cluster result, specifically includes：
Step B1：The second cluster centre is initialized as each first cluster centre in the first cluster result, and second is gathered
Class center clusters each first where current goal cluster centre as currently clustering as current goal cluster centre；
Specifically, the input data of the K-means algorithms by the first cluster result as an optimization carries out coordinate point set
Cluster operation.
Step B2：The coordinate points not traversed in traversal coordinate point set calculate the coordinate points traversed and each current goal
The distance between cluster centre therefrom selects minimum range, judges that minimum range is then whether no more than the radius parameter set
Perform step B3；Otherwise, the coordinate points of traversal are given up, performs step B4；
Wherein, the distance between the coordinate points traversed and each current goal cluster centre are calculated, preferably：According to traversal
The coordinate points that arrive and the coordinate of each current goal cluster centre, calculate the coordinate points and each current goal cluster centre that traverse it
Between Euclidean distance.
Step B3：It will be current poly- where the coordinate points cluster to the corresponding current goal cluster centre of minimum range of traversal
As currently clustering after cluster, the cluster centre currently to cluster is updated, and updated cluster centre is gathered as current goal
Class center performs step B4；
Step B4：Judge it is then return to step B2 with the presence or absence of the coordinate points that do not traverse in the coordinate point set；Otherwise
Each current cluster is clustered as second, using each current goal cluster centre as the second cluster centre, output second cluster and
Corresponding each second cluster centre, obtains the second cluster result.
In the present invention by cluster in the first cluster result containing 3 first and corresponding 3 the first cluster centres for into
Row explanation, based on the second cluster result that the first cluster result obtains, schematic diagram is as shown in Figure 3.
In the present invention, the DBSCAN algorithms after optimization substantially increase computational efficiency, but there are a defects, that is, lack
It is weary of overall importance, it can have the following problems：Some cluster set is not belonging to before certain point, but with the expansion of the cluster set
Exhibition, which is contained into, and the point has formd an individually cluster set before；To solve the defect, this hair
It is bright that K-means algorithms are optimized, and coordinate point set is carried out using the K-means algorithms after optimization primary global time
It goes through, coordinate points is returned in the clustering of radius parameter for gathering recently and meeting setting, more existing K-means algorithms carry out more
For secondary traversal, operation efficiency is not only increased, and be combined with the DBSCAN algorithms of optimization, solve the DBSCAN of optimization
Algorithm there are the defects of, improve clustering result quality.
Step 104：According to the second cluster result, the permanent residence of driver is predicted.
Specifically, each second cluster centre that will contain in the second cluster result, the permanent residence of the driver as prediction.
Embodiment two
According to the embodiment of the present invention, a kind of driver's permanent residence prediction meanss based on clustering algorithm are provided, such as Fig. 4 institutes
Show, including：
Collection module 201 forms data acquisition system for collecting the position data of driver；
First optimization module 202, for being optimized to DBSCAN algorithms；
First cluster module 203, for using the DBSCAN algorithms of the first optimization module 202 optimization to collection module 201
The data acquisition system of collection carries out cluster operation, obtains the first cluster result；
Second optimization module 204, for being optimized to K-means algorithms；
Second cluster module 205, it is excellent using second for the first cluster result obtained based on the first cluster module 203
Change the data acquisition system that the K-means algorithms that module 205 optimizes collect collection module 201 and carry out cluster operation, it is poly- to obtain second
Class result；
Prediction module 206 for the second cluster result obtained according to the second cluster module 205, predicts that driver's is resident
Ground.
According to the embodiment of the present invention, collection module 201 are specifically used for：Driver in preset time period is collected to be used
The coordinate points that report of client, form coordinate point set.
According to the embodiment of the present invention, the first optimization module 202, is specifically used for：Optimization the radius of neighbourhood meaning be：Certain
One current point adds in the basis for estimation of a certain cluster set；The meaning of Optimal Density threshold value is：A certain cluster set, which synthesizes, to cluster
Basis for estimation；
Accordingly, the first cluster module specifically includes：Select submodule, the first judging submodule, as submodule,
Two judging submodules, the first cluster submodule, the first newly-built submodule, the first computational submodule, third judging submodule, second
It clusters submodule, the second newly-built submodule, the 4th judging submodule, compare submodule and the first output sub-module, wherein：
Submodule is selected, for one pending coordinate of random selection in the coordinate point set collected in collection module 201
Point is as current point；
First judging submodule, for judging currently to cluster the quantity of set；
As submodule, when the quantity for judging currently to cluster set when the first judging submodule is zero, will select
Central point of the current point that submodule selects as first cluster set, and be current cluster set by first cluster set cooperation
It closes, triggering selection submodule；
Second judgment submodule when the quantity for judging currently to cluster set when the first judging submodule is 1, judges
Whether the current point and currently the distance between central point of cluster set for selecting submodule selection are not more than the first optimization module
The radius of neighbourhood of 202 optimizations；
First cluster submodule, for work as second judgment submodule judge to select the current point of submodule selection with it is current
When clustering the radius of neighbourhood of the distance between the central point of set no more than the optimization of the first optimization module 202, submodule will be selected
The current point cluster of selection updates the central point of current cluster set to current cluster set, the 4th selection submodule of triggering；
First newly-built submodule, for work as second judgment submodule judge to select the current point of submodule selection with it is current
When clustering the radius of neighbourhood of the distance between the central point of set more than the optimization of the first optimization module 202, cluster set is created；
First update submodule is additionally operable to that center of the current point that submodule selects as newly-built cluster set will be selected
Point, and newly-built cluster set cooperation is gathered for current cluster, the 4th selection submodule of triggering；
First computational submodule, for when the first judging submodule judges that the quantity for currently clustering set is more than 1, counting
The distance between the current point of selection submodule selection and the central point of each current cluster set are calculated, therefrom selects minimum range；
Third judging submodule, for judging whether the minimum range that the first computational submodule obtains optimizes no more than first
The radius of neighbourhood that module 202 optimizes；
Second cluster submodule, for working as the minimum range that third judging submodule judges that the first computational submodule obtains
No more than the optimization of the first optimization module 202 the radius of neighbourhood when, the current point cluster of submodule selection will be selected to minimum range
In corresponding current cluster set, the central point of the corresponding current cluster set of update minimum range, triggering the 4th judges submodule
Block；
Second newly-built submodule, for working as the minimum range that third judging submodule judges that the first computational submodule obtains
More than the first optimization module 202 optimization the radius of neighbourhood when, create cluster set；
Second cluster submodule, is additionally operable to select what the current point that submodule selects was created as the second newly-built submodule
The central point of set is clustered, and newly-built cluster set cooperation is gathered for current cluster, triggers the 4th judging submodule；
4th judging submodule, for judging whether also have untreated coordinate in the coordinate set of the collection of collection module 201
Point；
Select submodule, be additionally operable to when the 4th judging submodule judge collection module 201 collect coordinate set in also
When there are untreated coordinate points, a pending coordinate points conduct is randomly choosed in the coordinate point set collected in collection module 201
Current point；
Submodule is compared, for not locating when in the coordinate set that the 4th judging submodule judges collection module collection
When managing coordinate points, the density of each cluster set is compared with the density threshold that the first optimization module 202 optimizes successively, and will
Cluster set cooperation of the density not less than density threshold clusters for first, using corresponding central point as the first cluster centre；
First output sub-module, for export compare submodule obtain first cluster with corresponding first cluster in
The heart obtains the first cluster result.
According to the embodiment of the present invention, the second optimization module 204, is specifically used for：Radius parameter is set；
Accordingly, the second cluster module 205, specifically includes：Initialization submodule, traversal submodule, second calculate submodule
Block, third cluster submodule, gives up submodule, the 6th judging submodule and the second output sub-module at the 5th judging submodule,
In：
Initialization submodule, for initializing the first cluster result that the second cluster centre is the output of the first output sub-module
In each first cluster centre, and using the second cluster centre as current goal cluster centre, by current goal cluster centre institute
Each first cluster as currently clustering；
Submodule is traversed, for traversing the coordinate points not traversed in the coordinate point set of the collection of collection module 201；
Second computational submodule, for calculate the coordinate points that traverse of traversal submodule and each current goal cluster centre it
Between distance, therefrom select minimum range；
5th judging submodule, for judging whether the minimum range that the second computational submodule calculates optimizes no more than second
The radius parameter that module 204 is set；
Third clusters submodule, for judging the minimum range of the second computational submodule calculating when the 5th judging submodule
No more than the setting of the second optimization module 204 radius parameter when, the coordinate points cluster of submodule traversal will be traversed to minimum range
Current where corresponding current goal cluster centre cluster after as currently clustering, update the cluster centre currently to cluster,
And using updated cluster centre as current goal cluster centre, trigger the 6th judging submodule；
Give up submodule, for judging that the minimum range that the second computational submodule calculates is more than when the 5th judging submodule
During the radius parameter of the second optimization module 204 setting, the coordinate points for traversing submodule traversal are given up, triggering the 6th judges submodule
Block；
6th judging submodule, for judging to whether there is what is do not traversed in the coordinate point set of the collection of collection module 201
Coordinate points；
Submodule is traversed, for being deposited when in the coordinate point set that the 6th judging submodule judges the collection of collection module 201
In the coordinate points not traversed, the coordinate points not traversed in the coordinate point set that collection module 201 is collected are traversed；
Second output sub-module, for judging the coordinate point set of the collection of collection module 201 when the 6th judging submodule
In there is no during the coordinate points not traversed, each current cluster is clustered as second, using each current goal cluster centre as the
Two cluster centres, output second clusters and corresponding each second cluster centre, obtains the second cluster result.
According to the embodiment of the present invention, prediction module 206 are specifically used for：Second that second cluster module 205 is obtained
Each second cluster centre contained in cluster result, the permanent residence of the driver as prediction.
In the present invention, by the way that DBSCAN clustering algorithms and K-means algorithms are optimized and combined, i.e., based on optimization
DBSCAN algorithms carry out the position data of collection first time cluster, and the K- by the result of first time cluster as an optimization
The input data of means algorithms carries out second and clusters, not only increases computational efficiency, and ensure that the complete of cluster result
Office's property and high accuracy, and then correct guidance instruction can be provided for the reference person of data, promote industry and associated row
The development of industry.
The foregoing is only a preferred embodiment of the present invention, but protection scope of the present invention be not limited thereto,
Any one skilled in the art in the technical scope disclosed by the present invention, the change or replacement that can be readily occurred in,
It should be covered by the protection scope of the present invention.Therefore, protection scope of the present invention should be with the protection model of the claim
Subject to enclosing.
Claims (10)
1. a kind of driver's permanent residence Forecasting Methodology based on clustering algorithm, which is characterized in that including：
Step S1：The position data for collecting driver forms data acquisition system；
Step S2：DBSCAN algorithms are optimized, and the data acquisition system is clustered using the DBSCAN algorithms of optimization
Operation obtains the first cluster result；
Step S3：K-means algorithms are optimized, based on first cluster result, using the K-means algorithms pair of optimization
The data acquisition system carries out cluster operation and obtains the second cluster result；
Step S4：According to second cluster result, the permanent residence of the driver is predicted.
2. according to the method described in claim 1, it is characterized in that, the step S1, specially：It collects in preset time period and takes charge of
The coordinate points that client used in machine reports form coordinate point set.
3. according to the method described in claim 2, it is characterized in that, in the step S2, it is described DBSCAN algorithms are carried out it is excellent
Change, specifically include：
Optimization the radius of neighbourhood meaning be：A certain current point adds in the basis for estimation of a certain cluster set；
The meaning of Optimal Density threshold value is：A certain cluster set synthesizes the basis for estimation to cluster；
In the step S2, the DBSCAN algorithms using optimization carry out cluster operation to the data acquisition system, obtain first
Cluster result specifically includes：
Step A1：A pending coordinate points are randomly choosed in the coordinate point set as current point, judge current cluster
The quantity of set, the quantity of such as described current cluster set is zero, then performs step A2；Such as the quantity of the current cluster set
It is 1, then performs step A3；Quantity such as the current cluster set is more than 1, then performs step A4；
Step A2：It is using the central point that the current point is gathered as first cluster, and by first cluster set cooperation
Current cluster set, return to step A1；
Step A3：Judge the distance between the current point and the central point of the current cluster set whether no more than the neighbour
Domain radius is then to gather current point cluster to the current cluster, and update the central point of the current cluster set,
Perform step A5；Otherwise, cluster set is created, using the current point as the central point of newly-built cluster set, and will be created
Cluster set cooperation is gathered for current cluster, performs step A5；
Step A4：The distance between the current point and the central point of each current cluster set are calculated, therefrom selects minimum range,
Judge that the minimum range is then to correspond to current point cluster to the minimum range whether no more than the radius of neighbourhood
Current cluster set in, update the central point of the corresponding current cluster set of the minimum range, perform step A5；Otherwise,
Newly-built cluster set, using the central point that the current point is gathered as newly-built cluster, and by newly-built cluster set cooperation to work as
Preceding cluster set, performs step A5；
Step A5：Judge in the coordinate set whether to also have untreated coordinate points, be then return to step A1；Otherwise, step is performed
Rapid A6：
Step A6：The density of each cluster set of generation is compared, and density is not less than with the density threshold successively
The cluster set cooperation of the density threshold clusters for first, using corresponding central point as the first cluster centre, output described the
One clusters and corresponding first cluster centre, obtains the first cluster result.
4. according to the method described in claim 3, it is characterized in that, in the step S3, it is described K-means algorithms are carried out it is excellent
Change, specially：Radius parameter is set；
It is described based on first cluster result in the step S3, using the K-means algorithms of optimization to the data acquisition system
It carries out cluster operation and obtains the second cluster result, specifically include：
Step B1：The second cluster centre is initialized as each first cluster centre in first cluster result, and by described the
Two cluster centres cluster each first where the current goal cluster centre as current as current goal cluster centre
It clusters；
Step B2：The coordinate points not traversed in the coordinate point set are traversed, calculate the coordinate points traversed and each current goal
The distance between cluster centre therefrom selects minimum range, judges whether the minimum range is not more than the radius parameter of setting,
It is to perform step B3；Otherwise, the coordinate points of traversal are given up, performs step B4；
Step B3：It will be current poly- where the coordinate points cluster to the corresponding current goal cluster centre of the minimum range of traversal
As currently clustering after cluster, the cluster centre currently to cluster is updated, and updated cluster centre is gathered as current goal
Class center performs step B4；
Step B4：Judge it is then return to step B2 with the presence or absence of the coordinate points that do not traverse in the coordinate point set；It otherwise will be each
Current cluster clusters as second, using each current goal cluster centre as the second cluster centre, output described second cluster and
Corresponding each second cluster centre, obtains the second cluster result.
5. according to the method described in claim 4, it is characterized in that, the step S4, specially：By second cluster result
In each second cluster centre for containing, the permanent residence of the driver as prediction.
6. a kind of driver's permanent residence prediction meanss based on clustering algorithm, which is characterized in that including：
Collection module forms data acquisition system for collecting the position data of driver；
First optimization module, for being optimized to DBSCAN algorithms；
First cluster module, for being gathered using the DBSCAN algorithms of first optimization module optimization to the data acquisition system
Class operation obtains the first cluster result；
Second optimization module, for being optimized to K-means algorithms；
Second cluster module, for the first cluster result obtained based on first cluster module, using the second optimization module
The K-means algorithms of optimization carry out cluster operation to the data acquisition system that the collection module is collected, and obtain the first cluster result；
Prediction module for the second cluster result obtained according to second cluster module, predicts the permanent residence of the driver.
7. device according to claim 6, which is characterized in that the collection module is specifically used for：Collect preset time period
The coordinate points that client used in the corresponding driver of interior vehicle reports, form coordinate point set.
8. device according to claim 7, which is characterized in that
First optimization module, is specifically used for：Optimization the radius of neighbourhood meaning be：A certain current point adds in a certain cluster set
Basis for estimation；The meaning of Optimal Density threshold value is：A certain cluster set synthesizes the basis for estimation to cluster；
First cluster module, specifically includes：It selects submodule, the first judging submodule, judge son as submodule, second
Module, the first cluster submodule, the first newly-built submodule, the first computational submodule, third judging submodule, the second cluster submodule
Block, the 4th judging submodule, compares submodule and the first output sub-module at the second newly-built submodule；
The selection submodule, for randomly choosing a pending coordinate in the coordinate point set collected in the collection module
Point is as current point；
First judging submodule, for judging currently to cluster the quantity of set；
It is described to be used as submodule, it, will when the quantity for judging currently to cluster set when first judging submodule is zero
Central point of the current point that submodule is selected to select as first cluster set, and by first cluster set cooperation
Currently to cluster set, the selection submodule is triggered；
The second judgment submodule, the quantity that the current cluster set is judged for working as first judging submodule are
When 1, whether not the distance between central point that the current point of the selection submodule selection is gathered with the current cluster is judged
More than the radius of neighbourhood of first optimization module optimization；
The first cluster submodule judges the current of the selection submodule selection for working as the second judgment submodule
When the distance between point and the central point of the current cluster set are no more than the radius of neighbourhood that first optimization module optimizes,
By the current point cluster of the selection submodule selection to the current cluster set, and update in the current cluster set
Heart point, triggering the 4th selection submodule；
The first newly-built submodule judges the current of the selection submodule selection for working as the second judgment submodule
When the distance between point and the central point of the current cluster set are more than the radius of neighbourhood of first optimization module optimization, newly
Build cluster set；
The first update submodule is additionally operable to using the current point that submodule is selected to select as in newly-built cluster set
Heart point, and newly-built cluster set cooperation is gathered for current cluster, triggering the 4th selection submodule；
First computational submodule judges that the quantity of the current cluster set is big for working as first judging submodule
When 1, the distance between the current point of the selection submodule selection and the central point of each current cluster set, Cong Zhongxuan are calculated
Select minimum range；
The third judging submodule, for judging minimum range that first computational submodule obtains whether no more than described
The radius of neighbourhood of first optimization module optimization；
The second cluster submodule, judges what first computational submodule obtained for working as the third judging submodule
When minimum range is no more than the radius of neighbourhood that first optimization module optimizes, the current point of the selection submodule selection is gathered
In class to the corresponding current cluster set of the minimum range, the center of the corresponding current cluster set of the minimum range is updated
Point triggers the 4th judging submodule；
The second newly-built submodule judges what first computational submodule obtained for working as the third judging submodule
When minimum range is more than the radius of neighbourhood of first optimization module optimization, cluster set is created；
The second cluster submodule, is additionally operable to using the current point that submodule is selected to select as the described second newly-built submodule
The central point of cluster set that block creates, and newly-built cluster set cooperation is gathered for current cluster, triggering the described 4th judges
Submodule；
4th judging submodule, for judging whether also have untreated coordinate in the coordinate set of the collection module collection
Point；
The selection submodule is additionally operable to judge the coordinate set of the collection module collection when the 4th judging submodule
In also have untreated coordinate points when, the collection module collect coordinate point set in randomly choose a pending coordinate points
As current point；
The comparison submodule is judged for working as the 4th judging submodule in the coordinate set that the collection module is collected
When not having untreated coordinate points, the density of each cluster set and the density threshold of first optimization module optimization are carried out successively
It compares, and cluster set cooperation of the density not less than the density threshold is clustered for first, using corresponding central point as first
Cluster centre；
First output sub-module, for export it is described comparison submodule obtain first cluster with corresponding first cluster in
The heart obtains the first cluster result.
9. device according to claim 8, which is characterized in that
Second optimization module, is specifically used for：Radius parameter is set；
Second cluster module, specifically includes：Initialization submodule, traversal submodule, the second computational submodule, the 5th judge
Submodule, gives up submodule, the 6th judging submodule and the second output sub-module at third cluster submodule；
The initialization submodule, for initializing first cluster of second cluster centre for first output sub-module output
As a result each first cluster centre in, and using second cluster centre as current goal cluster centre, by the current mesh
Each first where mark cluster centre clusters as currently clustering；
The traversal submodule, for traversing the coordinate points not traversed in the coordinate point set of the collection module collection；
Second computational submodule, for calculating in the coordinate points and each current goal cluster that the traversal submodule traverses
The distance between heart therefrom selects minimum range；
5th judging submodule, for judging the minimum range of the second computational submodule calculating whether no more than described
The radius parameter of second optimization module setting；
The third clusters submodule, judges what second computational submodule calculated for working as the 5th judging submodule
When minimum range is no more than the radius parameter that second optimization module is set, the coordinate points of the traversal submodule traversal are gathered
Current where class to the corresponding current goal cluster centre of the minimum range cluster after as currently clustering, update this currently
The cluster centre to cluster, and using updated cluster centre as current goal cluster centre, triggering the described 6th judges submodule
Block；
It is described to give up submodule, for working as the minimum that the 5th judging submodule judges the second computational submodule calculating
When distance is more than the radius parameter of second optimization module setting, the coordinate points of the traversal submodule traversal are given up, are touched
Send out the 6th judging submodule described；
6th judging submodule, for judging to whether there is what is do not traversed in the coordinate point set of the collection module collection
Coordinate points；
The traversal submodule, for working as the coordinate point set that the 6th judging submodule judges the collection module collection
During the coordinate points that middle presence does not traverse, the coordinate points not traversed in the coordinate point set that the collection module is collected are traversed；
Second output sub-module, for working as the coordinate points that the 6th judging submodule judges the collection module collection
There is no during the coordinate points not traversed in set, each current cluster as second is clustered, each current goal cluster centre is made
For the second cluster centre, output described second clusters and corresponding each second cluster centre, obtains the second cluster result.
10. device according to claim 9, which is characterized in that the prediction module is specifically used for：Described second is gathered
Each second cluster centre contained in the second cluster result that generic module obtains, the permanent residence of the driver as prediction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711325708.4A CN108182446A (en) | 2017-12-13 | 2017-12-13 | A kind of driver's permanent residence Forecasting Methodology and device based on clustering algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711325708.4A CN108182446A (en) | 2017-12-13 | 2017-12-13 | A kind of driver's permanent residence Forecasting Methodology and device based on clustering algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108182446A true CN108182446A (en) | 2018-06-19 |
Family
ID=62546219
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711325708.4A Pending CN108182446A (en) | 2017-12-13 | 2017-12-13 | A kind of driver's permanent residence Forecasting Methodology and device based on clustering algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108182446A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109086323A (en) * | 2018-06-28 | 2018-12-25 | 上海中通吉网络技术有限公司 | The determination method and system of subscriber household and work address |
| CN109145989A (en) * | 2018-08-22 | 2019-01-04 | 深圳市东部公共交通有限公司 | Bus station distribution method, device and terminal |
| CN109257703A (en) * | 2018-10-09 | 2019-01-22 | 江苏满运软件科技有限公司 | Methods of exhibiting, device, electronic equipment, the storage medium of driver's accumulation point |
| CN109636020A (en) * | 2018-11-30 | 2019-04-16 | 广州亚美信息科技有限公司 | A kind of method and system for predicting destination-address |
| CN109978075A (en) * | 2019-04-04 | 2019-07-05 | 江苏满运软件科技有限公司 | Vehicle dummy location information identifying method, device, electronic equipment, storage medium |
| CN110298558A (en) * | 2019-06-11 | 2019-10-01 | 欧拉信息服务有限公司 | Vehicle resources dispositions method and device |
| CN111121803A (en) * | 2019-11-27 | 2020-05-08 | 北京中交兴路信息科技有限公司 | Method and device for acquiring common stop points of road |
| CN111372186A (en) * | 2019-12-17 | 2020-07-03 | 广东小天才科技有限公司 | Position calculation method under non-uniform positioning scene and terminal equipment |
| CN112380906A (en) * | 2020-10-19 | 2021-02-19 | 上汽通用五菱汽车股份有限公司 | Method for determining user address based on driving data |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104156489A (en) * | 2014-08-29 | 2014-11-19 | 北京嘀嘀无限科技发展有限公司 | Method for mining driver frequent parking points based on driver track |
| CN104167092A (en) * | 2014-07-30 | 2014-11-26 | 北京市交通信息中心 | Method and device for determining taxi pick-up and drop-off hot spot region center |
-
2017
- 2017-12-13 CN CN201711325708.4A patent/CN108182446A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104167092A (en) * | 2014-07-30 | 2014-11-26 | 北京市交通信息中心 | Method and device for determining taxi pick-up and drop-off hot spot region center |
| CN104156489A (en) * | 2014-08-29 | 2014-11-19 | 北京嘀嘀无限科技发展有限公司 | Method for mining driver frequent parking points based on driver track |
Non-Patent Citations (6)
| Title |
|---|
| ONAPA LIMWATTANAPIBOOL ET AL.: "Determination of the appropriate parameters for K‐means clustering using selection of region clusters based on density DBSCAN (SRCD‐DBSCAN)", 《EXPERT SYSTEMS》 * |
| 宓超等: "《装卸机器视觉及其应用》", 31 January 2016, 上海科学技术出版社 * |
| 张灿龙等: "一种改进DBSCAN密度聚类算法", 《数字技术与应用》 * |
| 李弼程等: "《模式识别原理与应用》", 29 February 2008, 西安电子科技大学出版社 * |
| 林强等: "《行为识别与智能计算》", 30 November 2016, 西安电子科技大学出版社 * |
| 龚键雅: "《对地观测数据处理与分析研究进展》", 31 December 2007, 武汉大学出版社 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109086323A (en) * | 2018-06-28 | 2018-12-25 | 上海中通吉网络技术有限公司 | The determination method and system of subscriber household and work address |
| CN109145989B (en) * | 2018-08-22 | 2021-07-06 | 深圳市东部公共交通有限公司 | Bus stop layout method and device and computer terminal |
| CN109145989A (en) * | 2018-08-22 | 2019-01-04 | 深圳市东部公共交通有限公司 | Bus station distribution method, device and terminal |
| CN109257703A (en) * | 2018-10-09 | 2019-01-22 | 江苏满运软件科技有限公司 | Methods of exhibiting, device, electronic equipment, the storage medium of driver's accumulation point |
| CN111866727B (en) * | 2018-10-09 | 2024-03-29 | 江苏满运软件科技有限公司 | Display method and device for driver gathering point, electronic equipment and storage medium |
| CN111866727A (en) * | 2018-10-09 | 2020-10-30 | 江苏满运软件科技有限公司 | Display method and device for driver gathering point, electronic equipment and storage medium |
| CN109636020A (en) * | 2018-11-30 | 2019-04-16 | 广州亚美信息科技有限公司 | A kind of method and system for predicting destination-address |
| CN109978075A (en) * | 2019-04-04 | 2019-07-05 | 江苏满运软件科技有限公司 | Vehicle dummy location information identifying method, device, electronic equipment, storage medium |
| CN109978075B (en) * | 2019-04-04 | 2021-09-28 | 江苏满运软件科技有限公司 | Vehicle false position information identification method and device, electronic equipment and storage medium |
| CN110298558A (en) * | 2019-06-11 | 2019-10-01 | 欧拉信息服务有限公司 | Vehicle resources dispositions method and device |
| CN111121803A (en) * | 2019-11-27 | 2020-05-08 | 北京中交兴路信息科技有限公司 | Method and device for acquiring common stop points of road |
| CN111121803B (en) * | 2019-11-27 | 2021-12-03 | 北京中交兴路信息科技有限公司 | Method and device for acquiring common stop points of road |
| CN111372186B (en) * | 2019-12-17 | 2021-08-24 | 广东小天才科技有限公司 | Position calculation method under non-uniform positioning scene and terminal equipment |
| CN111372186A (en) * | 2019-12-17 | 2020-07-03 | 广东小天才科技有限公司 | Position calculation method under non-uniform positioning scene and terminal equipment |
| CN112380906A (en) * | 2020-10-19 | 2021-02-19 | 上汽通用五菱汽车股份有限公司 | Method for determining user address based on driving data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108182446A (en) | A kind of driver's permanent residence Forecasting Methodology and device based on clustering algorithm | |
| Greene et al. | Tracking the evolution of communities in dynamic social networks | |
| Vidali et al. | A Deep Reinforcement Learning Approach to Adaptive Traffic Lights Management. | |
| CN106530715B (en) | Road net traffic state prediction technique based on fuzzy Markov process | |
| CN108269401A (en) | A kind of overpass traffic congestion Forecasting Methodology based on data-driven | |
| CN107610021A (en) | The comprehensive analysis method of environmental variance spatial and temporal distributions | |
| CN106447119A (en) | Short-term traffic flow prediction method and system based on convolutional neural network | |
| CN106875004A (en) | Composite mode neuronal messages processing method and system | |
| CN107317872B (en) | Scheduling method of multi-type tasks in space crowdsourcing | |
| CN101801011A (en) | WSN (Wireless Sensor Network) safety routing method based on credit evaluating mechanism | |
| CN109978215A (en) | Patrol management method and device | |
| CN105340311B (en) | The network equipment, network service prediction meanss and method | |
| CN111723973B (en) | Learning effect optimization method based on user behavior causal relationship in MOOC log data | |
| CN108510739A (en) | A kind of road traffic state recognition methods, system and storage medium | |
| CN109214559A (en) | The prediction technique and device of logistics business, readable storage medium storing program for executing | |
| Dahmardeh et al. | Agility evaluation in public sector using fuzzy logic | |
| CN110312206A (en) | Based on the improved mobile phone signaling data trip recognition methods of dynamic space threshold value | |
| CN107967513A (en) | Multirobot intensified learning collaboratively searching method and system | |
| CN104965974B (en) | A kind of method assessed based on coverage and optimize police strength resource deployment | |
| CN106777002A (en) | A kind of medical knowledge base system based on data mining | |
| CN109299778A (en) | A kind of calculation method of the RCRSS rescue map subregion based on cuckoo searching algorithm | |
| CN110414748A (en) | Photovoltaic power prediction technique | |
| CN110459050A (en) | A kind of short-term bus passenger flow prediction technique based on hybrid decision tree | |
| CN110138638A (en) | A kind of processing method and processing device of network flow | |
| CN109064750A (en) | Urban road network traffic estimation method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180619 |