CN109041218B - Method for predicting user position and intelligent hardware - Google Patents
Method for predicting user position and intelligent hardware Download PDFInfo
- Publication number
- CN109041218B CN109041218B CN201811117176.XA CN201811117176A CN109041218B CN 109041218 B CN109041218 B CN 109041218B CN 201811117176 A CN201811117176 A CN 201811117176A CN 109041218 B CN109041218 B CN 109041218B
- Authority
- CN
- China
- Prior art keywords
- target user
- user
- historical
- positioning data
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000000694 effects Effects 0.000 claims abstract description 117
- 238000004364 calculation method Methods 0.000 claims abstract description 84
- 230000000977 initiatory effect Effects 0.000 claims abstract description 42
- 230000007704 transition Effects 0.000 claims description 53
- 230000000007 visual effect Effects 0.000 claims description 7
- 238000004422 calculation algorithm Methods 0.000 abstract description 34
- 230000006399 behavior Effects 0.000 description 34
- 230000006870 function Effects 0.000 description 13
- 230000009471 action Effects 0.000 description 8
- 230000008901 benefit Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000007621 cluster analysis Methods 0.000 description 2
- 238000005034 decoration Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000012804 iterative process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000005192 partition Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W64/00—Locating users or terminals or network equipment for network management purposes, e.g. mobility management
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Telephonic Communication Services (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a method for predicting a user position and intelligent hardware, which comprises the following steps: performing clustering calculation according to the historical positioning data of the target user, and obtaining a target user activity area corresponding to the target user in each time period according to the clustering calculation result and the behavior rule data; clustering calculation is carried out on the historical positioning data of the neighbor users, and the neighbor user activity area corresponding to the target user positioning request moment of the neighbor users is obtained according to the clustering calculation result; counting the number of historical positioning data of a target user; when the number is larger than or equal to the preset number, predicting the position data of the target user at the moment of initiating the positioning request according to the activity area of the target user; and when the number is less than the preset number, predicting the position data of the target user at the current moment according to the activity area of the neighbor user. The invention carries out position prediction by using the historical positioning data of the user with minimum investment under limited resources, solves the problem of positioning drift, reduces the algorithm complexity and improves the positioning precision.
Description
Technical Field
The present invention relates to the field of information processing technologies, and in particular, to a method and an intelligent hardware for predicting a user location.
Background
With the popularization of mobile devices such as smart phones and tablets in recent years and the rapid development of positioning technologies, historical positioning data of various mobile devices are collected and utilized. Currently, the positioning technology mainly includes: GPS (global positioning system) positioning, Wi-Fi (wireless fidelity) positioning, base station positioning, and the like.
The positioning methods have certain errors. Generally, GPS positioning has high accuracy, but requires that GPS signals transmitted from GPS satellites be acquired during positioning, which causes high power consumption, and GPS positioning drifts when the GPS positioning system is located indoors, resulting in positioning errors. When the GPS signal is weak (e.g., indoors, building dense, under a bridge, etc.), accurate positioning cannot be achieved, and WiFi positioning or base station positioning may be used. If the positioning is realized by utilizing WiFi, extra hardware cost is needed, WiFi equipment is paved on a target building (uniformly deployed according to the characteristics of the building), the cost is huge, and the positioning precision is low. If the base station is needed to realize positioning, because the base station positioning belongs to ground positioning, the positioning accuracy depends on a signal tower, the signal tower needs to be additionally arranged near a target building, the cost is huge, and the positioning accuracy is low.
In the above case, the error of the positioning becomes large, causing a positioning drift. In order to correct positioning drift and improve positioning accuracy, map merchants often adopt ways of removing noise, binding roads, supplementing road shapes, rarefying and the like, but the methods are limited in application scenes and applicable to positioning scenes such as road positioning and navigation, but are not ideal in positioning effect for indoor scenes such as schools and offices. How to predict the position by using the historical positioning data of the user with the minimum investment under the limited resources is an urgent problem to be solved, the problem of positioning drift is solved, the algorithm complexity is reduced, and the positioning accuracy is improved.
Disclosure of Invention
The invention aims to provide a method for predicting a user position and intelligent hardware, which realize the position prediction by using the historical positioning data of a user with minimum investment under limited resources, solve the problem of positioning drift, reduce the algorithm complexity and improve the positioning precision.
The technical scheme provided by the invention is as follows:
the invention provides a method for predicting the position of a user, which comprises the following steps:
performing clustering calculation according to historical positioning data of the target user, and obtaining a target user activity area corresponding to the target user in each time period according to a clustering calculation result and behavior rule data;
clustering calculation is carried out on historical positioning data of neighbor users, and a neighbor user activity area corresponding to the time when the neighbor users initiate positioning requests is obtained according to clustering calculation results; the neighbor users are users whose behavior rule data similarity with the target user reaches a preset range and the number of historical positioning data reaches a preset number;
counting the number of historical positioning data of the target user;
when the number is larger than or equal to a preset number, predicting the position data of the target user at the moment of initiating the positioning request according to the activity area of the target user;
and when the number is less than the preset number, predicting the position data of the target user at the current moment according to the activity area of the neighbor user.
Further, before performing cluster calculation on the historical positioning data of more than one user to obtain a plurality of activity stay areas, the method includes:
acquiring historical positioning data of the target user in a preset time period and historical positioning data of the neighbor user in the preset time period;
the historical positioning data comprises any one or more of historical GPS positioning data, historical base station positioning data, historical WIFI positioning data, historical inertial navigation positioning data and historical visual positioning data.
Further, when the number is greater than or equal to a preset number, predicting, according to the target user activity area, the position data of the target user at the time of initiating the positioning request specifically includes:
when the number is larger than or equal to a preset number, calculating the transition probability of the position of the target user at each time point according to the activity area of the target user;
and obtaining the position data of the target user at the moment of initiating the positioning request according to the historical positioning data corresponding to the maximum transition probability.
Further, when the number is smaller than the preset number, predicting the position data of the target user at the current moment according to the neighbor user activity area specifically includes the steps of:
when the number is less than the preset number, calculating the transition probability of the positions of the neighbor users at each time point according to the activity areas of the neighbor users;
and obtaining the position data of the target user at the moment of initiating the positioning request according to the historical positioning data corresponding to the maximum transition probability.
Further, the transition probability is calculated according to the following formula:
wherein,
indicating that the user is at location l at any time tiAnd is in position at time t +1L placingjTotal number of times of (l), Σ count (l)i) Indicating that the user is at location l at any time tiThe total number of times.
The present invention also provides an intelligent hardware for predicting a user's location, comprising:
the target user activity area classification module is used for carrying out clustering calculation according to historical positioning data of a target user and obtaining a target user activity area corresponding to the target user in each time period according to a clustering calculation result and behavior rule data;
the neighbor user activity area classification module is used for carrying out clustering calculation on historical positioning data of neighbor users and obtaining a neighbor user activity area corresponding to the neighbor user at the moment when the target user initiates the positioning request according to a clustering calculation result; the neighbor users are users whose behavior rule data similarity with the target user reaches a preset range and the number of historical positioning data reaches a preset number;
the counting module is used for counting the number of the historical positioning data of the target user;
the user position prediction module is respectively connected with the statistical module, the target user activity area classification module and the neighbor user activity area classification module, and when the number is more than or equal to a preset number, the user position prediction module predicts the position data of the target user at the moment of initiating the positioning request according to the target user activity area; and when the number is less than the preset number, predicting the position data of the target user at the current moment according to the activity area of the neighbor user.
Further, the method also comprises the following steps:
the historical positioning data acquisition module is respectively connected with the counting module, the target user activity area classification module and the neighbor user activity area classification module and is used for acquiring historical positioning data of the target user in a preset time period and historical positioning data of the neighbor user in the preset time period;
the historical positioning data comprises any one or more of historical GPS positioning data, historical base station positioning data, historical WIFI positioning data, historical inertial navigation positioning data and historical visual positioning data.
Further, the user position prediction module is further configured to calculate, according to the target user activity area, a transition probability at a position where the target user is located at each time point when the number is greater than or equal to a preset number; and obtaining the position data of the target user at the moment of initiating the positioning request according to the historical positioning data corresponding to the maximum transition probability.
Further, the user position prediction module is further configured to calculate, according to the activity area of the neighbor user, a transition probability at a position where the neighbor user is located at each time point when the number is smaller than the preset number; and obtaining the position data of the target user at the moment of initiating the positioning request according to the historical positioning data corresponding to the maximum transition probability.
Further, the transition probability is calculated according to the following formula:
wherein,
indicating that the user is at location l at any time tiAnd at position l at time t +1jTotal number of times of (l), Σ count (l)i) Indicating that the user is at location l at any time tiThe total number of times.
By the method for predicting the user position and the intelligent hardware, the position prediction is realized by using the historical positioning data of the user with minimum investment under limited resources, the problem of positioning drift is solved, the algorithm complexity is reduced, and the positioning precision is improved.
Drawings
The above features, technical features, advantages and implementations of a method of predicting a user's location and intelligent hardware will be further described in the following detailed description of preferred embodiments in a clearly understandable manner, in conjunction with the accompanying drawings.
FIG. 1 is a flow diagram of one embodiment of a method of predicting a user's location of the present invention;
FIG. 2 is a schematic diagram of a clustering result obtained by clustering historical positioning data by using a K-means algorithm;
FIG. 3 is a flow chart of another embodiment of a method of predicting a user's location of the present invention;
FIG. 4 is a flow chart of another embodiment of a method of predicting a user's location of the present invention;
FIG. 5 is a block diagram illustrating one embodiment of the intelligent hardware for predicting a user's location in accordance with the present invention;
the reference numbers illustrate: the system comprises a target user activity area classification module 10, a neighbor user activity area classification module 20, a statistic module 30 and a user position prediction module 40.
Detailed Description
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following description will be made with reference to the accompanying drawings. It is obvious that the drawings in the following description are only some examples of the invention, and that for a person skilled in the art, other drawings and embodiments can be derived from them without inventive effort.
For the sake of simplicity, the drawings only schematically show the parts relevant to the present invention, and they do not represent the actual structure as a product. In addition, in order to make the drawings concise and understandable, components having the same structure or function in some of the drawings are only schematically illustrated or only labeled. In this document, "one" means not only "only one" but also a case of "more than one".
According to an embodiment of the present invention, as shown in fig. 1, a method for predicting a user location includes:
s100, performing clustering calculation according to historical positioning data of a target user, and obtaining a target user activity area corresponding to the target user in each time period according to a clustering calculation result and behavior rule data;
in particular, the process of grouping a set of physical or abstract objects into clusters of similar objects is referred to as cluster computation. The clusters generated by the clustering computation are a set of data objects that are similar to objects in the same cluster and different from objects in other clusters. In many applications, data objects in a cluster may be treated as a whole. The algorithm of the clustering calculation is many, such as hierarchical clustering algorithm, partition type clustering algorithm, K-means algorithm, etc.
The target user is a user who initiates a positioning request by using intelligent hardware with a positioning function, after clustering calculation is carried out according to historical positioning data of the target user to obtain a corresponding clustering calculation result, the clustering calculation result is divided according to behavior rule data of the target user obtained in advance to obtain a target user activity area corresponding to each time period of the target user. For example, if the target user is a student, the behavior data of the student on the study day is on-the-go, in the morning, in the afternoon, and on-the-go. The time periods corresponding to the behaviors are obtained according to the behavior rule data of the learning day, namely the time of the behavior rule data of the learning day is divided into the going-on-road time T1 ═ T1, T2, [ T2, [ T2, T3] of the morning time, the going-off-road time T3 ═ T3, T4] of the afternoon, the going-on-road time T4 ═ T4, T5] of the school, wherein T1T, T2, T3, T4 and T5 are all time points, and T1 < T2 < T3 < T4 < T5. And after the behavior time period segmentation is finished, respectively re-dividing the clustering calculation results according to the segmented behavior time period to obtain a target user activity area corresponding to each time period of the target user.
S200, clustering the historical positioning data of the neighbor users, and obtaining the neighbor user activity areas corresponding to the target user positioning request moments of the neighbor users according to clustering results; the neighbor users are users whose behavior rule data similarity with the target user reaches a preset range and the number of historical positioning data reaches a preset number;
specifically, the neighbor users include at least two users in the same area as the target user. For example, to obtain historical positioning data in smart phone watches of other students in the same school as the target user. When a target user initiates a positioning request by using intelligent hardware with a positioning function, acquiring historical positioning data of a neighbor user before the time of initiating the positioning request in real time, performing cluster calculation on the acquired historical positioning data of the neighbor user, and obtaining a neighbor user activity area corresponding to the time of initiating the positioning request by the neighbor user according to a cluster calculation result. Illustratively, the target user is at 18/8/21/2018: 00 when initiating a positioning request by using a smart phone watch with a positioning function, the smart phone watch acquires users A, B and C in the same area with a target user in 2018: year 8, month 21, day 18: 00, obtaining historical positioning data of neighbor users (A, B and C) for clustering calculation, and obtaining 18 of the neighbor users in 8 months and 21 days in 2018 according to clustering calculation results: 00 corresponds to the neighbor user activity area M. If the target user is 19 in 2018, 8, 21: 00 when initiating a positioning request by using a smart phone watch with a positioning function, similarly, the method can obtain that the neighbor user has the following information in 2018, 8, 21, 19: 00 corresponds to the neighbor user activity area N.
Preferably, the historical positioning data is clustered by using a K-means algorithm to obtain a corresponding clustering calculation result, and a plurality of historical positioning data are obtained, wherein the processing flow of clustering calculation by using the K-means algorithm is as follows:
1) and randomly extracting k observation points from all historical positioning data to serve as clustering center points, traversing the rest historical positioning data to find the clustering center point which is closest to the historical positioning data, and adding the clustering center points into the clustering calculation. Thus, there is an initial clustering calculation, which is an iterative process.
2) Each cluster calculation center has at least one historical positioning data, so that the central point (means) of each cluster calculation can be calculated to serve as a new cluster calculation center, then all observation points are traversed, the central point closest to the observation points is found, and the central point is added into the cluster calculation. Then operation 2) continues.
3) And repeating the step 2) until the clustering center points obtained by the two iterations are the same.
Let L bei＝(l1,l2,…,ln) The cluster computation result is represented by using the K-means algorithm as C (C), where Li is (Xi, Yi) for any given Li1,c2,…ck) And k is the input cluster calculation number, and the cluster center point of each cluster and the historical positioning data farthest from the cluster center point form a positioning area with the radius of r. As shown in fig. 2. The historical positioning data is gathered into 4 classes through a clustering algorithm, and the 4 classes are respectively represented by symbols A, B, C and D, wherein different symbols represent different classes. The euclidean distance is used here as a measure to evaluate how similar two historical positioning data features are. Because the historical positioning data is clustered and calculated by using the K-means algorithm, a training set does not need to be prepared, the principle is simple, and the method is easy to realize.
S300, counting the number of the historical positioning data of the target user;
s400, when the number is larger than or equal to a preset number, predicting the position data of the target user at the moment of initiating the positioning request according to the activity area of the target user;
s500, when the number is smaller than the preset number, predicting the position data of the target user at the current moment according to the activity area of the neighbor user.
Specifically, if the number of historical positioning data collected by the intelligent hardware used by the target user is larger than or equal to the preset number, the intelligent hardware is a device which is used for a long time, and a large amount of historical positioning data is stored in the intelligent hardware, so that the position data of the target user can be directly predicted according to the positioning data of the target user.
If the number of historical positioning data collected by the intelligent hardware used by the target user is smaller than the preset number, the intelligent hardware is a new device or a device which is not subjected to factory setting, so that enough historical positioning data are not stored in the intelligent hardware, and once the target user is in an indoor environment and the positioning data is inaccurate due to external environmental factors, the position data of the target user at the time of initiating the positioning request is predicted by combining with the activity area of a neighbor user.
The invention firstly carries out clustering calculation by using a clustering algorithm based on historical positioning data, and obtains different activity areas (namely the activity areas of target users or the activity areas of neighbor users) according to the historical positioning data of different users. And when the number of the historical positioning data of the target user is less than the preset number, predicting the position data of the target user at the current request initiating moment by using the activity area of the neighbor user. And when the number of the historical positioning data of the target user is larger than or equal to the preset number, predicting the position data of the target user at the positioning initiation moment by using the activity area of the target user. Therefore, the position prediction is carried out by using the user positioning data with minimum investment under limited resources, the problem of positioning drift is solved, the algorithm complexity is reduced, and the positioning accuracy, stability and instantaneity of the positioning system are improved.
Based on the foregoing embodiment, as shown in fig. 3, in this embodiment, the method further includes:
s010 obtains historical positioning data of the target user in a preset time period and historical positioning data of the neighbor user in the preset time period;
s100, performing clustering calculation according to historical positioning data of a target user, and obtaining a target user activity area corresponding to the target user in each time period according to a clustering calculation result and behavior rule data;
s200, clustering the historical positioning data of the neighbor users, and obtaining the neighbor user activity areas corresponding to the target user positioning request moments of the neighbor users according to clustering results; the neighbor users are users whose behavior rule data similarity with the target user reaches a preset range and the number of historical positioning data reaches a preset number;
s300, counting the number of the historical positioning data of the target user;
s400, when the number is larger than or equal to a preset number, predicting the position data of the target user at the moment of initiating the positioning request according to the activity area of the target user;
s500, when the number is smaller than the preset number, predicting the position data of the target user at the current moment according to the activity area of the neighbor user.
Specifically, the historical positioning data is acquired by using a positioning module of a smart phone or a smart phone watch. Because the coverage of intelligent hardware such as a smart phone watch or a smart phone is wide, the positioning data is generated in real time, the sampling period is long, and the positioning data of the user can be objectively recorded for a long time in a large scale. Moreover, as the number of users of the smart phone watch or the smart phone is large and the scale is large, the completeness and diversity of historical positioning data can be increased.
In addition, the historical positioning data comprises any one or more of historical GPS positioning data, historical base station positioning data, historical WIFI positioning data, historical inertial navigation positioning data and historical visual positioning data. If the acquired historical positioning data are of various types, clustering calculation results can be directly obtained from the different types of historical positioning data according to a clustering algorithm, then clustering calculation results corresponding to the different types of historical positioning data are calculated, then different activity areas are obtained, and after the position data are obtained according to activity area prediction corresponding to the different types of historical positioning data, mean value calculation is carried out on the position data, so that more excellent position data are obtained. Because different positioning data that the location type of difference obtained, and then obtain different position data, after the mean value is solved, can increase the variety and the difference of data, avoid data overfitting to make the position data who obtains more accurate.
Based on the foregoing embodiment, as shown in fig. 4, in this embodiment, the method further includes:
s100, performing clustering calculation according to historical positioning data of a target user, and obtaining a target user activity area corresponding to the target user in each time period according to a clustering calculation result and behavior rule data;
s200, clustering the historical positioning data of the neighbor users, and obtaining the neighbor user activity areas corresponding to the target user positioning request moments of the neighbor users according to clustering results; the neighbor users are users whose behavior rule data similarity with the target user reaches a preset range and the number of historical positioning data reaches a preset number;
s300, counting the number of the historical positioning data of the target user;
s410, when the number is larger than or equal to a preset number, calculating the transition probability of the target user at the position of each time point according to the activity area of the target user;
s420, according to the historical positioning data corresponding to the maximum transition probability, obtaining the position data of the target user at the moment of initiating the positioning request;
specifically, clustering calculation is performed on historical positioning data of a target user by using a clustering algorithm, and each clustering calculation result is a target user activity area. And time information is provided based on historical positioning data, so that the positions of the target users at various time points can be obtained. The time point does not refer to a time stamp, but refers to a point of the same time corresponding to each preset time interval. For example, if the preset time interval is 1 day, if the acquisition of the historical positioning data is 30 days, then the acquisition results in 08: 00 this time point is the location of the target user. According to the activity area of the target user, calculating the transition probability of the target user at the position of each time point, wherein the transition probability is calculated according to the following formula:
wherein,
indicating that the user is at location l at any time tiAnd at position l at time t +1jTotal number of times of (l), Σ count (l)i) Indicating that the user is at location l at any time tiThe total number of times.
And after calculating the transition probability of the target user at each time point, comparing the transition probabilities, and thus obtaining the position data of the target user at the moment of initiating the positioning request according to the historical positioning data corresponding to the maximum transition probability.
Illustratively, the target user activity area of the target user includes location H, location P, and location Q, if the user is at 18/8/21/2018: 00: 00 initiates a location request, then if historical location data for 100 target users is acquired. Assume that the target user is at 17: 59: 59 is in position H, 18: 00: 00 are located at position Q65 times, the target user is at 17: 59: 59 is in position Q, 18: 00: 00 are located at position P18 times, the target user is at 17: 59: 59 is in position Q, 18: 00: 00 is located at the position H17 times, then it can be calculated that the target user is located at the time of initiating the location request, i.e. 18 in 2018, 8, 21 th: the location of 00 is the location Q, i.e. the location data is the location Q.
S510, when the number is smaller than the preset number, calculating the transition probability of the positions of the neighbor users at each time point according to the activity areas of the neighbor users;
s520, according to the historical positioning data corresponding to the maximum transition probability, the position data of the target user at the moment of initiating the positioning request is obtained.
Specifically, the neighbor users are users whose behavior habits are similar to those of the target user and whose number of historical positioning data reaches a preset number, and which are obtained by screening the behavior rule data of the target user. Behavior data includes action range, action time, and the like. For example, the target user is a student in a class-one class of a certain teaching building in a school, and the neighbor user is a student who is at the same desk as the target user and has a preset number of historical positioning data. Of course, other neighbor user identities may be selected, such as a table that is different from the target user on the same class, but students whose historical positioning data count reaches a predetermined number may also be used as neighbor users. For example, the target user is a student in a class and a class of a certain teaching building in a school, wherein although the neighbor user has students in the same school as the target user, the students are not the neighbor user because the class is different, so that some students and the target user are not in the same teaching building and class, that is, the matching degree of the action time or the action range is not high. For another example, although some students are in the same class as the target user, the number of the historical positioning data does not reach the preset number, and therefore the students are not neighbor users.
Firstly, clustering calculation is carried out on historical positioning data of neighbor users by using a clustering algorithm, each clustering calculation result is a neighbor user activity area, and the neighbor user activity areas comprise neighbor user activity areas. And time information is acquired based on historical positioning data, so that the positions of the neighbor users at all time points can be acquired. According to the activity area of the neighbor user, the position of the neighbor user at each time point calculates the transition probability, and the transition probability is calculated according to the following formula:
wherein,
indicating that the user is at location l at any time tiAnd at position l at time t +1jTotal number of times of (l), Σ count (l)i) Indicating that the user is at location l at any time tiThe total number of times.
And after calculating the transition probability of the position of the neighbor user at each time point, comparing the magnitude of each transition probability, and thus obtaining the position data of the target user at the moment of initiating the positioning request according to the historical positioning data corresponding to the maximum transition probability.
In this embodiment, the inaccurate problem of present indoor location has been solved, gather the function of intelligent hardware and calculate through cluster analysis, after clustering target user's historical positioning data and neighbor user's historical positioning data respectively and obtaining corresponding clustering result, carry out ingenious calculation and obtain the transition probability, thereby obtain the target user and initiate the position data of location request moment according to the historical positioning data that the transition probability is the biggest, indoor user location accuracy has been improved, utilize daily necessities promptly the smart mobile phone simultaneously, the function of smart mobile phone wrist-watch is used to the location in, low consumption has, convenient advantage.
The invention utilizes the historical positioning data collected by massive intelligent hardware to divide the activity area by combining with the behavior rule data (the work and rest time rule of the user), thereby effectively solving the problem of positioning drift. The method for matching the activity area with the time can greatly reduce the calculation amount of the algorithm and improve the positioning precision, the stability and the real-time property. Auxiliary hardware acquisition equipment is not required to be additionally arranged for auxiliary positioning, and map merchants such as Baidu maps and Google maps do not need road network data, so that the capital cost can be effectively reduced.
According to an embodiment of the present invention, as shown in fig. 5, an intelligent hardware for predicting a user's location includes:
the target user activity area classification module 10 is used for performing clustering calculation according to historical positioning data of a target user, and obtaining a target user activity area corresponding to the target user in each time period according to a clustering calculation result and behavior rule data;
in particular, the process of grouping a set of physical or abstract objects into clusters of similar objects is referred to as cluster computation. The clusters generated by the clustering computation are a set of data objects that are similar to objects in the same cluster and different from objects in other clusters. In many applications, data objects in a cluster may be treated as a whole. The algorithm of the clustering calculation is many, such as hierarchical clustering algorithm, partition type clustering algorithm, K-means algorithm, etc.
The target user is a user who initiates a positioning request by using intelligent hardware with a positioning function, after clustering calculation is carried out according to historical positioning data of the target user to obtain a corresponding clustering calculation result, the clustering calculation result is divided according to behavior rule data of the target user obtained in advance to obtain a target user activity area corresponding to each time period of the target user. For example, if the target user is a student, the behavior data of the student on the study day is on-the-go, in the morning, in the afternoon, and on-the-go. The time periods corresponding to the behaviors are obtained according to the behavior rule data of the learning day, namely the time of the behavior rule data of the learning day is divided into the going-on-road time T1 ═ T1, T2, [ T2, [ T2, T3] of the morning time, the going-off-road time T3 ═ T3, T4] of the afternoon, the going-on-road time T4 ═ T4, T5] of the school, wherein T1T, T2, T3, T4 and T5 are all time points, and T1 < T2 < T3 < T4 < T5. And after the behavior time period segmentation is finished, respectively re-dividing the clustering calculation results according to the segmented behavior time period to obtain a target user activity area corresponding to each time period of the target user.
The neighbor user activity area classification module 20 is configured to perform clustering calculation on historical positioning data of neighbor users, and obtain a neighbor user activity area corresponding to the neighbor user at the time when the target user initiates the positioning request according to a clustering calculation result; the neighbor users are users whose behavior rule data similarity with the target user reaches a preset range and the number of historical positioning data reaches a preset number;
specifically, the neighbor users include at least two users in the same area as the target user. For example, to obtain historical positioning data in smart phone watches of other students in the same school as the target user. When a target user initiates a positioning request by using intelligent hardware with a positioning function, acquiring historical positioning data of a neighbor user before the time of initiating the positioning request in real time, performing cluster calculation on the acquired historical positioning data of the neighbor user, and obtaining a neighbor user activity area corresponding to the time of initiating the positioning request by the neighbor user according to a cluster calculation result. Illustratively, the target user is at 18/8/21/2018: 00 when initiating a positioning request by using a smart phone watch with a positioning function, the smart phone watch acquires users A, B and C in the same area with a target user in 2018: year 8, month 21, day 18: 00, obtaining historical positioning data of neighbor users (A, B and C) for clustering calculation, and obtaining 18 of the neighbor users in 8 months and 21 days in 2018 according to clustering calculation results: 00 corresponds to the neighbor user activity area M. If the target user is 19 in 2018, 8, 21: 00 when initiating a positioning request by using a smart phone watch with a positioning function, similarly, the method can obtain that the neighbor user has the following information in 2018, 8, 21, 19: 00 corresponds to the neighbor user activity area N.
Preferably, the historical positioning data is clustered by using a K-means algorithm to obtain a corresponding clustering calculation result, and a plurality of historical positioning data are obtained, wherein the processing flow of clustering calculation by using the K-means algorithm is as follows:
1) and randomly extracting k observation points from all historical positioning data to serve as clustering center points, traversing the rest historical positioning data to find the clustering center point which is closest to the historical positioning data, and adding the clustering center points into the clustering calculation. Thus, there is an initial clustering calculation, which is an iterative process.
2) Each cluster calculation center has at least one historical positioning data, so that the central point (means) of each cluster calculation can be calculated to serve as a new cluster calculation center, then all observation points are traversed, the central point closest to the observation points is found, and the central point is added into the cluster calculation. Then operation 2) continues.
3) And repeating the step 2) until the clustering center points obtained by the two iterations are the same.
Let L bei＝(l1,l2,…,ln) The cluster computation result is represented by using the K-means algorithm as C (C), where Li is (Xi, Yi) for any given Li1,c2,…ck) And k is the input cluster calculation number, and the cluster center point of each cluster and the historical positioning data farthest from the cluster center point form a positioning area with the radius of r. As shown in fig. 2. The historical positioning data is gathered into 4 classes through a clustering algorithm, and the 4 classes are respectively represented by symbols A, B, C and D, wherein different symbols represent different classes. The euclidean distance is used here as a measure to evaluate how similar two historical positioning data features are. Due to the use of K-means algorithmThe method carries out cluster calculation on the historical positioning data, so that a training set does not need to be prepared, the principle is simple, and the method is easy to realize.
A counting module 30, configured to count the number of historical positioning data of the target user;
a user position prediction module 40, connected to the statistics module 30, the target user activity region classification module 10 and the neighbor user activity region classification module 20, respectively, for predicting the position data of the target user at the time of initiating the positioning request according to the target user activity region when the number is greater than or equal to a preset number; and when the number is less than the preset number, predicting the position data of the target user at the current moment according to the activity area of the neighbor user.
Specifically, if the number of historical positioning data collected by the intelligent hardware used by the target user is larger than or equal to the preset number, the intelligent hardware is a device which is used for a long time, and a large amount of historical positioning data is stored in the intelligent hardware, so that the position data of the target user can be directly predicted according to the positioning data of the target user.
If the number of historical positioning data collected by the intelligent hardware used by the target user is smaller than the preset number, the intelligent hardware is a new device or a device which is not subjected to factory setting, so that enough historical positioning data are not stored in the intelligent hardware, and once the target user is in an indoor environment and the positioning data is inaccurate due to external environmental factors, the position data of the target user at the moment of initiating the positioning request cannot be predicted by combining the activity area of a neighbor user.
The invention firstly carries out clustering calculation by using a clustering algorithm based on historical positioning data, and obtains different activity areas (namely the activity areas of target users or the activity areas of neighbor users) according to the historical positioning data of different users. And when the number of the historical positioning data of the target user is less than the preset number, predicting the position data of the target user at the current request initiating moment by using the activity area of the neighbor user. And when the number of the historical positioning data of the target user is larger than or equal to the preset number, predicting the position data of the target user at the positioning initiation moment by using the activity area of the target user. Therefore, the position prediction is carried out by using the user positioning data with minimum investment under limited resources, the problem of positioning drift is solved, the algorithm complexity is reduced, and the positioning accuracy, stability and instantaneity of the positioning system are improved.
Based on the foregoing embodiment, in this embodiment, the method further includes:
a historical positioning data acquiring module, which is respectively connected with the counting module 30, the target user activity area classifying module 10 and the neighbor user activity area classifying module 20, and is used for acquiring historical positioning data of the target user in a preset time period and historical positioning data of the neighbor user in the preset time period;
the historical positioning data comprises any one or more of historical GPS positioning data, historical base station positioning data, historical WIFI positioning data, historical inertial navigation positioning data and historical visual positioning data.
Specifically, the historical positioning data is acquired by using a positioning module of a smart phone or a smart phone watch. Because the coverage of intelligent hardware such as a smart phone watch or a smart phone is wide, the positioning data is generated in real time, the sampling period is long, and the positioning data of the user can be objectively recorded for a long time in a large scale. Moreover, as the number of users of the smart phone watch or the smart phone is large and the scale is large, the completeness and diversity of historical positioning data can be increased.
In addition, the historical positioning data comprises any one or more of historical GPS positioning data, historical base station positioning data, historical WIFI positioning data, historical inertial navigation positioning data and historical visual positioning data. If the acquired historical positioning data are of various types, clustering calculation results can be directly obtained from the different types of historical positioning data according to a clustering algorithm, then clustering calculation results corresponding to the different types of historical positioning data are calculated, then different activity areas are obtained, and after the position data are obtained according to activity area prediction corresponding to the different types of historical positioning data, mean value calculation is carried out on the position data, so that more excellent position data are obtained. Because different positioning data that the location type of difference obtained, and then obtain different position data, after the mean value is solved, can increase the variety and the difference of data, avoid data overfitting to make the position data who obtains more accurate.
Based on the foregoing embodiment, in this embodiment, the method further includes:
the user position prediction module 40 is further configured to calculate, according to the target user activity area, a transition probability at a position where the target user is located at each time point when the number is greater than or equal to a preset number; and obtaining the position data of the target user at the moment of initiating the positioning request according to the historical positioning data corresponding to the maximum transition probability.
Specifically, clustering calculation is performed on historical positioning data of a target user by using a clustering algorithm, and each clustering calculation result is a target user activity area. And time information is provided based on historical positioning data, so that the positions of the target users at various time points can be obtained. The time point does not refer to a time stamp, but refers to a point of the same time corresponding to each preset time interval. For example, if the preset time interval is 1 day, if the acquisition of the historical positioning data is 30 days, then the acquisition results in 08: 00 this time point is the location of the target user. According to the activity area of the target user, calculating the transition probability of the target user at the position of each time point, wherein the transition probability is calculated according to the following formula:
wherein,
indicating that the user is at location l at any time tiAnd at position l at time t +1jTotal number of times of (l), Σ count (l)i) Indicating that the user is at location l at any time tiThe total number of times.
And after calculating the transition probability of the target user at each time point, comparing the transition probabilities, and thus obtaining the position data of the target user at the moment of initiating the positioning request according to the historical positioning data corresponding to the maximum transition probability.
Illustratively, the target user activity area of the target user includes location H, location P, and location Q, if the user is at 18/8/21/2018: 00: 00 initiates a location request, then if historical location data for 100 target users is acquired. Assume that the target user is at 17: 59: 59 is in position H, 18: 00: 00 are located at position Q65 times, the target user is at 17: 59: 59 is in position Q, 18: 00: 00 are located at position P18 times, the target user is at 17: 59: 59 is in position Q, 18: 00: 00 is located at the position H17 times, then it can be calculated that the target user is located at the time of initiating the location request, i.e. 18 in 2018, 8, 21 th: the location of 00 is the location Q, i.e. the location data is the location Q.
The user position prediction module 40 is further configured to calculate, according to the activity area of the neighbor user, a transition probability at a position where the neighbor user is located at each time point when the number is smaller than the preset number; and obtaining the position data of the target user at the moment of initiating the positioning request according to the historical positioning data corresponding to the maximum transition probability.
The transition probability is calculated according to the following formula:
wherein,
indicating that the user is at location l at any time tiAnd at position l at time t +1jTotal number of times of (l), Σ count (l)i) Indicating that the user is at location l at any time tiThe total number of times.
Specifically, the neighbor users are users whose behavior habits are similar to those of the target user and whose number of historical positioning data reaches a preset number, and which are obtained by screening the behavior rule data of the target user. Behavior data includes action range, action time, and the like. For example, the target user is a student in a class-one class of a certain teaching building in a school, and the neighbor user is a student who is at the same desk as the target user and has a preset number of historical positioning data. Of course, other neighbor user identities may be selected, such as a table that is different from the target user on the same class, but students whose historical positioning data count reaches a predetermined number may also be used as neighbor users. For example, the target user is a student in a class and a class of a certain teaching building in a school, wherein although the neighbor user has students in the same school as the target user, the students are not the neighbor user because the class is different, so that some students and the target user are not in the same teaching building and class, that is, the matching degree of the action time or the action range is not high. For another example, although some students are in the same class as the target user, the number of the historical positioning data does not reach the preset number, and therefore the students are not neighbor users.
Firstly, clustering calculation is carried out on historical positioning data of neighbor users by using a clustering algorithm, each clustering calculation result is a neighbor user activity area, and the neighbor user activity areas comprise neighbor user activity areas. And time information is acquired based on historical positioning data, so that the positions of the neighbor users at all time points can be acquired. According to the activity area of the neighbor user, the position of the neighbor user at each time point calculates the transition probability, and the transition probability is calculated according to the following formula:
wherein,
indicating that the user is at location l at any time tiAnd at position l at time t +1jTotal number of times of (l), Σ count (l)j) Indicating that the user is at location l at any time tiThe total number of times.
And after calculating the transition probability of the position of the neighbor user at each time point, comparing the magnitude of each transition probability, and thus obtaining the position data of the target user at the moment of initiating the positioning request according to the historical positioning data corresponding to the maximum transition probability.
In this embodiment, the inaccurate problem of present indoor location has been solved, gather the function of intelligent hardware and calculate through cluster analysis, after clustering target user's historical positioning data and neighbor user's historical positioning data respectively and obtaining corresponding clustering result, carry out ingenious calculation and obtain the transition probability, thereby obtain the target user and initiate the position data of location request moment according to the historical positioning data that the transition probability is the biggest, indoor user location accuracy has been improved, utilize daily necessities promptly the smart mobile phone simultaneously, the function of smart mobile phone wrist-watch is used to the location in, low consumption has, convenient advantage.
The invention utilizes the historical positioning data collected by massive intelligent hardware to divide the activity area by combining with the behavior rule data (the work and rest time rule of the user), thereby effectively solving the problem of positioning drift. The method for matching the activity area with the time can greatly reduce the calculation amount of the algorithm and improve the positioning precision, the stability and the real-time property. Auxiliary hardware acquisition equipment is not required to be additionally arranged for auxiliary positioning, and map merchants such as Baidu maps and Google maps do not need road network data, so that the capital cost can be effectively reduced.
It should be noted that the above embodiments can be freely combined as necessary. The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (10)
1. A method of predicting a user's location, comprising the steps of:
performing clustering calculation according to historical positioning data of the target user, and obtaining a target user activity area corresponding to the target user in each time period according to a clustering calculation result and behavior rule data;
clustering calculation is carried out on historical positioning data of neighbor users, and a neighbor user activity area corresponding to the time when the neighbor users initiate positioning requests is obtained according to clustering calculation results; the neighbor users are users whose behavior rule data similarity with the target user reaches a preset range and the number of historical positioning data reaches a preset number;
counting the number of historical positioning data of the target user;
when the number is larger than or equal to a preset number, predicting the position data of the target user at the moment of initiating the positioning request according to the activity area of the target user;
and when the number is less than the preset number, predicting the position data of the target user at the current moment according to the activity area of the neighbor user.
2. The method of claim 1, wherein clustering historical positioning data of more than one user to obtain a plurality of activity stay areas comprises:
acquiring historical positioning data of the target user in a preset time period and historical positioning data of the neighbor user in the preset time period;
the historical positioning data comprises any one or more of historical GPS positioning data, historical base station positioning data, historical WIFI positioning data, historical inertial navigation positioning data and historical visual positioning data.
3. The method according to claim 1, wherein when the number is greater than or equal to a preset number, predicting the location data of the target user at the time of initiating the positioning request according to the target user activity area specifically comprises:
when the number is larger than or equal to a preset number, calculating the transition probability of the position of the target user at each time point according to the activity area of the target user;
and obtaining the position data of the target user at the moment of initiating the positioning request according to the historical positioning data corresponding to the maximum transition probability.
4. The method according to claim 1, wherein when the number is less than a preset number, predicting the location data of the target user at the current time according to the neighbor user activity area specifically comprises the steps of:
when the number is less than the preset number, calculating the transition probability of the positions of the neighbor users at each time point according to the activity areas of the neighbor users;
and obtaining the position data of the target user at the moment of initiating the positioning request according to the historical positioning data corresponding to the maximum transition probability.
5. The method of predicting a user's location according to claim 3 or 4, wherein said transition probability is calculated according to the following formula:
wherein,
indicating that the user is at location l at any time tiAnd at position l at time t +1jTotal number of times of (l), Σ count (l)i) Indicating that the user is at location l at any time tiThe total number of times.
6. Intelligent hardware for predicting a user's location, comprising:
the target user activity area classification module is used for carrying out clustering calculation according to historical positioning data of a target user and obtaining a target user activity area corresponding to the target user in each time period according to a clustering calculation result and behavior rule data;
the neighbor user activity area classification module is used for carrying out clustering calculation on historical positioning data of neighbor users and obtaining a neighbor user activity area corresponding to the neighbor user at the moment when the target user initiates the positioning request according to a clustering calculation result; the neighbor users are users whose behavior rule data similarity with the target user reaches a preset range and the number of historical positioning data reaches a preset number;
the counting module is used for counting the number of the historical positioning data of the target user;
the user position prediction module is respectively connected with the statistical module, the target user activity area classification module and the neighbor user activity area classification module, and when the number is more than or equal to a preset number, the user position prediction module predicts the position data of the target user at the moment of initiating the positioning request according to the target user activity area; and when the number is less than the preset number, predicting the position data of the target user at the current moment according to the activity area of the neighbor user.
7. The intelligent hardware for predicting the location of a user of claim 6, further comprising:
the historical positioning data acquisition module is respectively connected with the counting module, the target user activity area classification module and the neighbor user activity area classification module and is used for acquiring historical positioning data of the target user in a preset time period and historical positioning data of the neighbor user in the preset time period;
the historical positioning data comprises any one or more of historical GPS positioning data, historical base station positioning data, historical WIFI positioning data, historical inertial navigation positioning data and historical visual positioning data.
8. The intelligent hardware for predicting the location of a user of claim 6, wherein:
the user position prediction module is further configured to calculate a transition probability according to the target user activity area and the position of the target user at each time point when the number is greater than or equal to a preset number; and obtaining the position data of the target user at the moment of initiating the positioning request according to the historical positioning data corresponding to the maximum transition probability.
9. The intelligent hardware for predicting the location of a user of claim 6, wherein:
the user position prediction module is further used for calculating the transition probability of the positions of the neighbor users at each time point according to the activity areas of the neighbor users when the number is smaller than the preset number; and obtaining the position data of the target user at the moment of initiating the positioning request according to the historical positioning data corresponding to the maximum transition probability.
10. Intelligent hardware to predict user position according to claim 8 or 9, characterized in that said transition probability is calculated according to the following formula:
wherein,
indicating that the user is at location l at any time tiAnd at position l at time t +1jTotal number of times of (l), Σ count (l)i) Indicating that the user is at location l at any time tiThe total number of times.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811117176.XA CN109041218B (en) | 2018-09-25 | 2018-09-25 | Method for predicting user position and intelligent hardware |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811117176.XA CN109041218B (en) | 2018-09-25 | 2018-09-25 | Method for predicting user position and intelligent hardware |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109041218A CN109041218A (en) | 2018-12-18 |
| CN109041218B true CN109041218B (en) | 2020-08-11 |
Family
ID=64617931
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811117176.XA Active CN109041218B (en) | 2018-09-25 | 2018-09-25 | Method for predicting user position and intelligent hardware |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109041218B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110569321B (en) * | 2019-07-23 | 2023-04-11 | 平安科技（深圳）有限公司 | Grid division processing method and device based on urban map and computer equipment |
| CN110611880B (en) * | 2019-09-26 | 2021-05-04 | 广东小天才科技有限公司 | Household WiFi prediction method and device, electronic equipment and storage medium |
| CN113537548B (en) * | 2020-04-21 | 2024-04-23 | 杭州海康威视数字技术股份有限公司 | Recommendation method, device and equipment for driving route |
| CN112968934B (en) * | 2021-01-29 | 2023-02-17 | 北京骑胜科技有限公司 | Travel positioning method, travel positioning device, storage medium and computer program product |
| CN114007186B (en) * | 2021-10-21 | 2022-12-09 | 腾讯科技（深圳）有限公司 | Positioning method and related product |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102066969A (en) * | 2008-05-16 | 2011-05-18 | 苹果公司 | Location determination |
| CN104239556A (en) * | 2014-09-25 | 2014-12-24 | 西安理工大学 | Density clustering-based self-adaptive trajectory prediction method |
| CN105222768A (en) * | 2014-06-30 | 2016-01-06 | 奇点新源国际技术开发(北京)有限公司 | A kind of positioning track Forecasting Methodology and device |
| CN105409306A (en) * | 2014-06-11 | 2016-03-16 | 华为技术有限公司 | Method and apparatus for predicting location of mobile terminal |
| CN105893537A (en) * | 2016-03-31 | 2016-08-24 | 百度在线网络技术（北京）有限公司 | Determination method and device for geographical information points |
| CN106488405A (en) * | 2016-12-29 | 2017-03-08 | 电子科技大学 | A kind of position predicting method merging individuality and neighbour's movement law |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103731916B (en) * | 2014-01-14 | 2017-07-07 | 上海河广信息科技有限公司 | A kind of customer location forecasting system and method based on wireless network |
| CN105933857B (en) * | 2015-11-25 | 2019-05-14 | 中国银联股份有限公司 | A kind of mobile terminal locations prediction technique and device |
| JP6240235B2 (en) * | 2016-02-19 | 2017-11-29 | ヤフー株式会社 | Determination device, determination method, and determination program |
-
2018
- 2018-09-25 CN CN201811117176.XA patent/CN109041218B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102066969A (en) * | 2008-05-16 | 2011-05-18 | 苹果公司 | Location determination |
| CN105409306A (en) * | 2014-06-11 | 2016-03-16 | 华为技术有限公司 | Method and apparatus for predicting location of mobile terminal |
| CN105222768A (en) * | 2014-06-30 | 2016-01-06 | 奇点新源国际技术开发(北京)有限公司 | A kind of positioning track Forecasting Methodology and device |
| CN104239556A (en) * | 2014-09-25 | 2014-12-24 | 西安理工大学 | Density clustering-based self-adaptive trajectory prediction method |
| CN105893537A (en) * | 2016-03-31 | 2016-08-24 | 百度在线网络技术（北京）有限公司 | Determination method and device for geographical information points |
| CN106488405A (en) * | 2016-12-29 | 2017-03-08 | 电子科技大学 | A kind of position predicting method merging individuality and neighbour's movement law |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109041218A (en) | 2018-12-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109041218B (en) | Method for predicting user position and intelligent hardware | |
| CN110264709B (en) | Method for predicting traffic flow of road based on graph convolution network | |
| CN107610464B (en) | A kind of trajectory predictions method based on Gaussian Mixture time series models | |
| CN110414732B (en) | Travel future trajectory prediction method and device, storage medium and electronic equipment | |
| KR102097426B1 (en) | Analysis method of fluidized population information capable of providing real-time fluidized population data by pcell algorithm | |
| CN106709606B (en) | Personalized scene prediction method and device | |
| CN108834077B (en) | Tracking area division method and device based on user movement characteristics and electronic equipment | |
| CN107688906B (en) | Multi-method fused transmission line meteorological element downscaling analysis system and method | |
| CN104408203A (en) | Method for predicting path destination of moving object | |
| CN103929719B (en) | The optimization method and optimization device of location information | |
| CN106339716A (en) | Mobile trajectory similarity matching method based on weighted Euclidean distance | |
| CN112543419B (en) | User trajectory prediction method and device based on density clustering | |
| CN104793224A (en) | GPS positioning rectification method and device | |
| CN110399445B (en) | Method, device and equipment for processing interest points | |
| CN110326323B (en) | Method and device for acquiring emission probability, transition probability and sequence positioning | |
| CN112579922B (en) | Interest point recommendation method based on user track | |
| CN109874104B (en) | User position positioning method, device, equipment and medium | |
| CN106951455A (en) | A kind of similar track analysis system and its analysis method | |
| CN112738729A (en) | Method and system for distinguishing visiting hometown visitor by mobile phone signaling data | |
| CN111782980A (en) | Map interest point mining method, device, equipment and storage medium | |
| CN107133689B (en) | Position marking method | |
| CN104202817B (en) | The storage of the many focuses received signals fingerprint map in the large-scale interior space and matching process | |
| Zhao et al. | Urban crowd flow forecasting based on cellular network | |
| CN107577727B (en) | Group movement behavior characteristic analysis method | |
| CN113326877A (en) | Model training method, data processing method, device, apparatus, storage medium, and program |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |