CN106294485B - Determine the method and device in significant place - Google Patents

Abstract

The invention provides a method and a device for determining a prominent place, and relates to the technical field of the mobile Internet. It is invented to solve the problem of low recognition accuracy of prominent locations in the prior art. The method includes step 1: determining the determining attribute of the salient location according to the sample position trajectory and step 2: determining the salient location in the location trajectory to be processed according to the determining attribute of the salient location. Among them, step 1 specifically includes: determining the potential salient locations on the sample location trajectory; obtaining the value of the first attribute set corresponding to each potentially salient location to obtain the attribute information table; according to the attribute information table, using the preset feature selection algorithm to select the Selecting a salient location to determine the attribute from an attribute set; obtaining the threshold range of the salient location to determine the attribute. Step 2 specifically includes: determining at least one potential salient location on the location track to be processed; and determining a potentially salient location whose value of a determining attribute of a salient location is within a threshold range as a salient location.

Description

Method and device for determining prominent locations

technical field

The present invention relates to the technical field of mobile Internet, in particular to a method and device for determining a prominent location.

Background technique

With the prevalence of mobile Internet, the behavior analysis of mobile terminal users has become the focus of research. Among them, in the user behavior analysis, the user's location analysis is of great significance. Based on the location of the mobile terminal user, the location that the user visits more frequently, that is, the prominent location, can be obtained to provide the user with more intelligent information when the user is in a prominent location. Serve.

The prior art provides a method for determining salient locations. In the implementation process of this method, firstly, preprocessing is performed on the obtained location trajectory, and the points that are obviously not salient locations are eliminated, and then clustering algorithm is used for the remaining points After processing, each point is marked as a noise point and a salient point, and the point marked as a salient point is the determined salient point. The existing method for determining salient places may have the problem of misjudging non-prominent places as salient places in the process of identifying salient places, so the existing technology has low recognition accuracy for salient places.

Contents of the invention

The present invention provides a method and device for determining prominent locations to solve the problem of low recognition of prominent locations in the prior art.

To achieve the above object, the present invention adopts the following technical solutions:

In a first aspect, the present invention provides a method for determining a salient location, the method comprising: determining a salient location determining attribute according to a sample location trajectory and determining a salient location in the location trajectory to be processed according to the salient location determining attribute;

Wherein, the determination of the determining attribute of the salient location according to the sample position track specifically includes:

Obtaining at least one potentially significant location on the trajectory of the sample location through a preset clustering algorithm;

Acquiring the value of the first attribute set corresponding to each of the potentially significant locations to obtain an attribute information table, the first attribute set includes a preset number of conditional attributes and a decision attribute;

According to the attribute information table, using a preset feature selection algorithm to select a prominent place from the first attribute set to determine an attribute;

Obtaining the threshold range of the determining attribute of the prominent location;

The determining the salient locations in the location track to be processed according to the salient location determination attributes specifically includes:

Obtaining at least one potentially salient location on the location track to be processed through the preset clustering algorithm;

Potentially notable places whose value of the salient place determining attribute is within the threshold range are determined as salient places.

In combination with the first aspect, in the first implementation manner of the first aspect, according to the attribute information table, using a preset feature selection algorithm to select a prominent location from the first attribute set to determine an attribute, specifically includes:

Using a preset feature selection algorithm to calculate the value of the P-norm with tolerance capability of each conditional attribute in the first attribute set;

The condition attribute corresponding to the maximum value among all the calculated values of the P-norm with tolerance capability is determined as the salient location determining attribute.

In combination with the first implementation of the first aspect, in the second implementation of the first aspect, the use of a preset feature selection algorithm to calculate the P with tolerance capability of each conditional attribute in the first attribute set Norm values, including:

Using the first conditional attribute in the first attribute set as the attribute to be evaluated, and determining the set formed by the remaining conditional attributes as the second attribute set, the first conditional attribute is any conditional attribute;

Using a user attribute selection game algorithm based on fuzzy entropy, sequentially judge whether the attribute to be evaluated wins in the co-tenancy game of each target subset in the second attribute set, and obtain a judgment result corresponding to each target subset , the target subset includes at least two conditional attributes;

According to all the judgment results, the value of the P-norm with tolerance capability of the property to be evaluated is calculated.

With reference to the first aspect, in a third implementation manner of the first aspect, the preset clustering algorithm includes three parameters: distance threshold, point threshold and time threshold;

The obtaining at least one potential salient location on the sample position track or the position track to be processed through a preset clustering algorithm specifically includes:

Using an unmarked location in the location track as a starting point, the marking includes a potentially salient location or a noise point;

Find a target point whose distance from the starting point is less than or equal to the distance threshold;

If the number of the target points found is greater than or equal to the point threshold, and the time interval between each target point and the starting point is greater than the time threshold, then the starting point and all the targets points are determined as a cluster, and the starting point is marked as a potentially salient location;

After all points in the cluster are marked, other unmarked places are marked in sequence.

In combination with the first aspect or any one of the first implementation, the second implementation, and the third implementation of the first aspect, in the fourth implementation of the first aspect, the first The attribute set includes velocity, acceleration, the number of records of potentially significant locations in the sample location track, the number of days that the potentially prominent locations appear in the sample location track, the dwell time interval and orientation of the potentially prominent locations Standard deviation of variation.

In the second aspect, the present invention also provides a device for determining a significant location, and the device includes:

A processing module, configured to obtain at least one potentially significant location on the sample location trajectory through a preset clustering algorithm;

An acquisition module, configured to acquire the value of the first attribute set corresponding to each of the potentially significant locations, and obtain an attribute information table, where the first attribute set includes a preset number of conditional attributes and a decision attribute;

The processing module is further configured to use a preset feature selection algorithm to select a prominent place from the first attribute set to determine an attribute according to the attribute information table;

The obtaining module is also used to obtain the threshold range of the determining attribute of the prominent place;

The processing module is further configured to obtain at least one potentially significant location on the track of the position to be processed through the preset clustering algorithm;

The processing module is further configured to determine, as a salient place, a potential notable place whose value of the salient place determining attribute is within the threshold range.

With reference to the second aspect, in the first implementation manner of the second aspect, the acquisition module is specifically configured to use a preset feature selection algorithm to calculate the P with tolerance capability of each conditional attribute in the first attribute set. the value of the norm;

The condition attribute corresponding to the maximum value among all the calculated values of the P-norm with tolerance capability is determined as the salient location determining attribute.

In combination with the first implementation of the second aspect, in the second implementation of the second aspect, the acquisition module is also specifically used for:

Using the first conditional attribute in the first attribute set as the attribute to be evaluated, and determining the set formed by the remaining conditional attributes as the second attribute set, the first conditional attribute is any conditional attribute;

Using a user attribute selection game algorithm based on fuzzy entropy, sequentially judge whether the attribute to be evaluated wins in the co-tenancy game of each target subset in the second attribute set, and obtain a judgment result corresponding to each target subset , the target subset includes at least two conditional attributes;

According to all the judgment results, the value of the P-norm with tolerance capability of the property to be evaluated is calculated.

With reference to the second aspect, in a third implementation manner of the second aspect, the preset clustering algorithm includes three parameters: distance threshold, point threshold and time threshold;

The processing module is specifically used for:

Using an unmarked location in the location track as a starting point, the marking includes a potentially salient location or a noise point;

Find a target point whose distance from the starting point is less than or equal to the distance threshold;

If the number of the target points found is greater than or equal to the point threshold, and the time interval between each target point and the starting point is greater than the time threshold, then the starting point and all the targets points are determined as a cluster, and the starting point is marked as a potentially salient location;

After all points in the cluster are marked, other unmarked places are marked in sequence.

In combination with the second aspect or any one of the first implementation, the second implementation, and the third implementation of the second aspect, in the fourth implementation of the second aspect, the acquisition module The acquired first attribute set includes velocity, acceleration, the number of records of potentially significant locations in the sample location track, the number of days that the potentially prominent locations appear in the sample location track, the potential significant location Dwell intervals and standard deviations of azimuth changes.

The method and device for determining salient locations provided by the embodiments of the present invention use the sample location trajectory to obtain potential salient locations through a clustering algorithm, and then select the decisive attributes of the salient locations and their threshold ranges through a feature selection algorithm, and then select the location to be processed Trajectories are processed, and the potential salient locations in the potential salient locations in the position track to be processed are determined as the salient locations whose value of the decisive attribute of the salient locations meets the threshold range, and the points selected by the clustering algorithm are directly used as the salient locations in the prior art. Compared with the prominent locations, the present invention can reduce the misjudgment rate of the prominent locations due to the adoption of the feature selection algorithm after the clustering algorithm, thereby improving the accuracy of determining the prominent locations.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. Those skilled in the art can also obtain other drawings based on these drawings without creative work.

FIG. 1 is a schematic flowchart of a method for determining a prominent location provided by an embodiment of the present invention;

FIG. 2 is a schematic flow chart of a method for determining potential salient locations on a location trajectory through a preset clustering algorithm provided by an embodiment of the present invention;

Fig. 3 is a schematic flowchart of a method for selecting prominent locations and determining attributes according to the attribute information table provided by an embodiment of the present invention using a preset feature selection algorithm;

FIG. 4 is a schematic flowchart of a method for calculating the value of the P-norm with tolerance capability for each conditional attribute provided by an embodiment of the present invention;

Fig. 5 is a schematic structural diagram of a device for determining prominent locations provided by an embodiment of the present invention;

FIG. 6 is a schematic structural diagram of another device for determining a prominent location provided by an embodiment of the present invention.

Detailed ways

The following will clearly and completely describe the technical solution in this embodiment with reference to the drawings in this embodiment. Obviously, the described embodiment is only a part of the embodiments of the present invention, but not all of them. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

An embodiment of the present invention provides a method for determining a salient location, the method comprising: determining a salient location determining attribute according to a sample location trajectory and determining a salient location in a location trajectory to be processed according to the salient location determining attribute.

Wherein, the sample location trajectory is a location trajectory that has already determined which locations are salient locations. The position track to be processed is a position track for which no significant point has been determined. The sample position track and the pending position track, as a kind of position track, both include the position information of the mobile terminal at various time points within a period of time or the route it has traveled. Generally speaking, the existing mobile terminal has a GPS sensor, which can receive and record the continuous GPS longitude and latitude coordinate points of the user's mobile terminal to form a GPS track, which is the location track referred to in the embodiment of the present invention.

The "determining the salient location decision attribute according to the sample position trajectory" specifically includes the following steps 101 to 104; the "determining the salient location in the pending location trajectory according to the salient location determination attribute" specifically includes the following steps 105 and 104. 106.

As shown in Figure 1, the method for determining a prominent location provided by the embodiment of the present invention specifically includes:

101: Obtain at least one potentially significant location on the sample location track by using a preset clustering algorithm.

Among them, the basic purpose of the clustering algorithm is to form a cluster of points that meet specific conditions. Common clustering algorithms include the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm.

Points on the location trajectory can be divided into noise points and potential salient points according to their probabilities of becoming salient points.

102: Obtain the value of the first attribute set corresponding to each of the potentially significant locations to obtain an attribute information table, where the first attribute set includes a preset number of condition attributes and a decision attribute.

Wherein, the values of each conditional attribute and decision attribute in the first attribute set can be directly read by analyzing the sample position track or obtained after a certain operation. For example: when the number of records of the potential location in the location track is used as a condition attribute, it can be obtained by directly searching the number of occurrences of the potential location. Another example: when speed is a conditional attribute, after determining a potential significant point, for a potential significant point, the path can be obtained by obtaining a very small path including the potentially significant point and the time corresponding to the path, and the speed can be calculated The formula obtains the speed of the mobile terminal in this short period of time and takes the speed as the speed of the potential salient point.

As an optional implementation, the first attribute set may include velocity, acceleration, the number of records of potentially significant locations in the sample location trajectory, and the number of days that the potentially significant locations appear in the sample location trajectory , the dwell time interval of the potentially significant location and the standard deviation of the orientation change, a total of 6 conditional attributes. The decision-making attribute referred to refers to whether the potential prominent place is a prominent place.

Taking the example where the sample location trajectory includes 10 potentially salient locations, and the first attribute set table includes the above-mentioned 6 conditional attributes and 1 decision attribute as an example, Table 1 below gives the attribute information table corresponding to the potentially salient locations.

Table 1: Attribute Information Table

Among them, in the column of whether it is a prominent place, "1" means that the potential prominent place is a prominent place, and "0" means that the potential prominent place is not a prominent place.

The above Table 1 only shows the realization form of the attribute information table, and the value of each conditional attribute can be obtained after obtaining according to the actual situation.

103: According to the attribute information table, use a preset feature selection algorithm to select a prominent place from the first attribute set to determine an attribute.

Wherein, the preset feature selection algorithm referred to may be a P-norm with tolerance capability.

104: Obtain a threshold range of the determining attribute of the prominent location.

In a specific implementation of this step, by obtaining the value of the prominent location decision attribute corresponding to multiple locations (including both non-significant locations and prominent locations); if a certain interval of the value of the prominent location decision attribute contains If the number of salient locations is the largest, then this interval is determined as the threshold range of the salient locations determining attributes. Taking the conditional attribute of speed as an example to determine the salient location, the determination process is as follows:

Obtain the velocity values of multiple locations on the sample position trajectory;

The interval containing the speed value with the largest number of prominent locations is determined as the threshold range of the conditional attribute of speed.

105: Obtain at least one potentially significant location on the location track to be processed by using the preset clustering algorithm.

106: Determine the potential salient places whose value of the salient place determining attribute is within the threshold range as the salient places.

In the method for determining a salient location provided by the embodiment of the present invention, the sample location trajectory is obtained through a clustering algorithm to obtain a potential salient location, and then the determining attribute of the salient location and its threshold range are selected through a feature selection algorithm, and then the location trajectory to be processed is processed. processing, determining the potential salient places whose value of the determining attribute of the salient places in the position trajectory to be processed meets the threshold range as the salient places, which is different from the prior art in which the points selected by the clustering algorithm are directly used as the salient places In contrast, the present invention can reduce the misjudgment rate of prominent locations because the feature selection algorithm is used after the clustering algorithm, thereby improving the accuracy of determining prominent locations.

It should be noted that, in order to improve the selection accuracy of the salient location determining attribute and the threshold range of the determining attribute in the embodiment of the present invention, it is possible to obtain multiple sample position trajectories and use the same method from step 101 to step 104 to obtain the same method as Each sample position trajectory corresponds to a determining attribute and the threshold value range of the determining attribute, and then after analysis and processing, the final significant location determining attribute and its threshold value range are obtained.

As an implementation of the acquisition of potentially significant locations on the location trajectory described in step 101 and step 105: the referred preset clustering algorithm is a variant of the DBSCAN algorithm, and the preset clustering algorithm includes a distance threshold and a point threshold and time threshold three parameters.

The obtaining of at least one potential salient location on the sample position track or the position track to be processed by the preset clustering algorithm, as shown in FIG. 2 , specifically includes:

201: Use an unmarked location in the location track as a starting point.

Wherein, the marking includes marking as a potentially salient location or a noise point.

202: Search for a target point whose distance to the starting point is less than or equal to the distance threshold.

As an example, the value of the distance threshold may be 100m.

203: If the number of the found target points is greater than or equal to the point threshold, and the time interval between each target point and the starting point is greater than the time threshold, then set the starting point and all The target point is determined as a cluster, and the starting point is marked as a potentially salient location.

As an example, the point threshold may be 5; the time threshold may be 10 minutes.

If the above conditions are not satisfied, the starting point is marked as a noise point.

204: After all the points in the cluster are marked, mark other unmarked places in sequence.

Steps 201 to 203 are repeatedly executed, and all points in the cluster are processed sequentially. When all points in the cluster are marked (it can be said that the cluster is fully expanded), other unmarked points are marked.

The clustering algorithm used in the embodiment of the present invention is a variant of the DBSCAN algorithm, which not only considers the distance between points and points in the trajectory, but also the distance between points and points is the distance on the trajectory rather than the distance between two points. Direct distance, which is closer to the real situation in real life. In addition, in the clustering process, the minimum time is used instead of the minimum number of regions to determine clusters, which can prevent the user equipment from having to stay here for a short period of time due to a special reason, resulting in misjudgment of significant locations.

As a specific implementation of step 103 "according to the attribute information table, use a preset feature selection algorithm to select a prominent place to determine an attribute from the first attribute set", as shown in Figure 3, the process specifically includes:

301: Use a preset feature selection algorithm to calculate a P-norm value with tolerance capability for each conditional attribute in the first attribute set.

302: Determine the conditional attribute corresponding to the maximum value among all the calculated values of the P-norm with tolerance capability as the salient location determining attribute.

Wherein, as shown in FIG. 4, in the above step 301, the specific process of calculating the value of the P norm with tolerance capability for each conditional attribute is:

401: Use a first conditional attribute in the first attribute set as an attribute to be evaluated, and determine a set formed by other conditional attributes as a second attribute set, where the first conditional attribute is any conditional attribute.

402: Using a user attribute selection game algorithm based on fuzzy entropy, sequentially judge whether the attribute to be evaluated wins in the co-tenancy game of each target subset in the second attribute set, and obtain the value corresponding to each target subset. As a result of the judgment, the target subset includes at least two condition attributes.

The basic principle of the implementation process of this step is to use the user attribute selection game method based on fuzzy entropy, calculate the conditional mutual information, and judge whether the attribute to be evaluated wins in the shared rental game of the attribute subset.

403: According to all the judgment results, calculate the value of the P-norm with tolerance capability of the attribute to be evaluated.

By repeating the above step 401 to step 403, the value of the P-norm with tolerance capability corresponding to each condition attribute can be obtained.

Assuming that there are N-1 conditional attributes Ni and 1 decision attribute D in the first attribute set, and any conditional attribute Ni in the potential location information table is selected as the attribute to be evaluated for evaluation, the remaining N-2 conditional attributes constitute The second set of attributes. The second attribute set corresponds to a plurality of subsets, select any attribute subset Si in the second attribute set that contains at least two conditional attributes, to evaluate the relationship between the attribute subset Si and Degree of information sharing for final decision attributes. Its calculation formula is as follows:

MI(Si;D|Ni)＝E(Si|Ni)-E(Si|D,Ni) (1)

Among them, the conditional entropy E(Si|Ni) refers to the entropy of the attribute subset Si when Ni is known, and E(Si|D,Ni) indicates that when the conditional attribute Ni and the decision attribute D to be evaluated exist at the same time, the attribute The entropy of the subset Si.

If the value of MI(Si; D|Ni) in formula (1) is greater than 0, it means that when the attribute Ni to be evaluated exists, the information sharing between the attribute subset Si and the decision attribute D is increased, and it is said that Ni is in the attribute subset Si wins the co-tenancy game.

According to the calculation method of formula (1), it can be obtained sequentially whether the property Ni corresponding to the property to be evaluated wins in the rent-sharing game of each target attribute subset Si.

In addition, the calculation method of conditional entropy E(Si|Ni) is:

E(Si|Ni)＝-(∑log ₂ (|[xi]|/n))/n (2)

Where, |[xi]|=∑r _ij , r _ij is the similarity relationship between any two condition attribute values x _i and x _j in the Si set, and n is the number of condition attributes in Si. The calculation formula of r _ij is shown in the following formula (3):

Wherein, in formula (3), || _xi -x _j || is a P norm with tolerance capability.

Calculate the value of P-norm (Banzhaf) of the attribute to be evaluated according to whether the attribute to be evaluated wins in the rent-sharing game of each target attribute subset.

In the formula (1), when MI(Si; D|Ni) is greater than 0, that is, when the conditional attribute Ni wins in the co-tenancy game of the attribute subset Si, count Δ _i (Si) = 1, otherwise Δ _i (Si) = 0.

Then the Banzhaf value calculation method of the condition attribute Ni is as follows:

Finally, the Banzhaf value corresponding to each conditional attribute is obtained. The higher the Banzhaf value, the greater its contribution to the attribute subset. The condition attribute corresponding to the highest value of Banzhaf value is selected as the most important attribute for determining the salient location, that is, the salient location determining attribute.

Since the potentially salient locations are obtained through a clustering algorithm, there may be cases where samples are sparse, and the P-norm with tolerance capability used in feature extraction is more able to deal with the case of sparse samples than the traditional P-norm. The information loss caused by discretization of equivalence relation samples is avoided.

In addition, before step 101 and step 105, since the position trajectory acquired by the GPS sensor will drift from time to time, it is necessary to remove these drift points that are obviously noise points, such as drifting from Beijing to Shanghai in an instant, which is obviously wrong , such GPS latitude and longitude coordinate points need to be eliminated.

As an implementation of the above method, an embodiment of the present invention also provides a device for determining a prominent location, as shown in Figure 5, the device includes:

A processing module 501, configured to obtain at least one potentially significant location on the sample location trajectory through a preset clustering algorithm;

An acquisition module 502, configured to acquire the value of the first attribute set corresponding to each of the potential significant locations, and obtain an attribute information table, the first attribute set includes a preset number of conditional attributes and a decision attribute;

The processing module 501 is further configured to use a preset feature selection algorithm to select a prominent place from the first attribute set to determine an attribute according to the attribute information table;

The acquiring module 502 is further configured to acquire the threshold range of the determining attribute of the prominent place;

The processing module 501 is further configured to use the preset clustering algorithm to obtain at least one potential salient location on the track of the location to be processed;

The processing module 501 is further configured to determine, as a salient place, a potential notable place whose value of the salient place determining attribute is within the threshold range.

Further, the acquisition module 502 is specifically configured to use a preset feature selection algorithm to calculate the value of the P-norm with tolerance capability of each conditional attribute in the first attribute set;

The condition attribute corresponding to the maximum value among all the calculated values of the P-norm with tolerance capability is determined as the salient location determining attribute.

Further, the obtaining module 502 is also specifically used for:

Using the first conditional attribute in the first attribute set as the attribute to be evaluated, and determining the set formed by the remaining conditional attributes as the second attribute set, the first conditional attribute is any conditional attribute;

Using a user attribute selection game algorithm based on fuzzy entropy, sequentially judge whether the attribute to be evaluated wins in the co-tenancy game of each target subset in the second attribute set, and obtain a judgment result corresponding to each target subset , the target subset includes at least two conditional attributes;

According to all the judgment results, the value of the P-norm with tolerance capability of the attribute to be evaluated is calculated.

Further, the preset clustering algorithm includes three parameters: distance threshold, points threshold and time threshold; then the processing module 501 is specifically used for:

Using an unmarked location in the location track as a starting point, the marking includes a potentially salient location or a noise point;

Find a target point whose distance from the starting point is less than or equal to the distance threshold;

If the number of the target points found is greater than or equal to the point threshold, and the time interval between each target point and the starting point is greater than the time threshold, then the starting point and all the targets points are determined as a cluster, and the starting point is marked as a potentially salient location;

After all points in the cluster are marked, other unmarked places are marked in sequence.

Further, the first attribute set acquired by the acquisition module 502 includes velocity, acceleration, the number of records of potential salient points in the sample position track, and the number of records of the potential salient points in the sample position track , the dwell time interval and the standard deviation of the change in orientation of said potentially significant location.

The device for determining a salient location provided by the embodiment of the present invention uses a clustering algorithm to obtain a potential salient location from the sample location trajectory, and then selects the decisive attribute of the salient location and its threshold range through a feature selection algorithm, and then performs the processing on the location trajectory to be processed. processing, determining the potential salient places whose value of the determining attribute of the salient places in the position trajectory to be processed meets the threshold range as the salient places, which is different from the prior art in which the points selected by the clustering algorithm are directly used as the salient places In contrast, the present invention can reduce the misjudgment rate of prominent locations because the feature selection algorithm is used after the clustering algorithm, thereby improving the accuracy of determining prominent locations.

As an implementation of the above method, an embodiment of the present invention also provides a device for determining a significant location, as shown in FIG. . in:

Processor 601, configured to obtain at least one potentially significant location on the sample location trajectory through a preset clustering algorithm;

Acquiring the value of the first attribute set corresponding to each of the potentially significant locations to obtain an attribute information table, the first attribute set includes a preset number of conditional attributes and a decision attribute;

According to the attribute information table, using a preset feature selection algorithm to select a prominent place from the first attribute set to determine an attribute;

Obtaining the threshold range of the determining attribute of the prominent location;

Obtaining at least one potentially salient location on the location track to be processed through the preset clustering algorithm;

Potentially notable places whose value of the salient place determining attribute is within the threshold range are determined as salient places.

Further, the processor 601 is specifically configured to use a preset feature selection algorithm to calculate the value of the P-norm with tolerance capability of each conditional attribute in the first attribute set;

The condition attribute corresponding to the maximum value among all the calculated values of the P-norm with tolerance capability is determined as the salient location determining attribute.

Further, the processor 601 is also specifically configured to:

Using the first conditional attribute in the first attribute set as the attribute to be evaluated, and determining the set formed by the remaining conditional attributes as the second attribute set, the first conditional attribute is any conditional attribute;

Using a user attribute selection game algorithm based on fuzzy entropy, sequentially judge whether the attribute to be evaluated wins in the co-tenancy game of each target subset in the second attribute set, and obtain a judgment result corresponding to each target subset , the target subset includes at least two conditional attributes;

According to all the judgment results, the value of the P-norm with tolerance capability of the attribute to be evaluated is calculated.

Further, the processor 601 is specifically configured to:

Using an unmarked location in the location track as a starting point, the marking includes a potentially salient location or a noise point;

Find a target point whose distance from the starting point is less than or equal to the distance threshold;

If the number of the target points found is greater than or equal to the point threshold, and the time interval between each target point and the starting point is greater than the time threshold, then the starting point and all the targets points are determined as a cluster, and the starting point is marked as a potentially salient location;

After all points in the cluster are marked, other unmarked places are marked in sequence.

Further, the first attribute set acquired by the processor 601 includes velocity, acceleration, the number of records of potential salient points in the sample position track, and the number of records of the potential salient points in the sample position track. , the dwell time interval and the standard deviation of the change in orientation of said potentially significant location.

The memory 602 is used for storing the programs used in the execution process of the processor 601 .

The device for determining a salient location provided by the embodiment of the present invention uses a clustering algorithm to obtain a potential salient location from the sample location trajectory, and then selects the decisive attribute of the salient location and its threshold range through a feature selection algorithm, and then performs the processing on the location trajectory to be processed. processing, determining the potential salient places whose value of the determining attribute of the salient places in the position trajectory to be processed meets the threshold range as the salient places, which is different from the prior art in which the points selected by the clustering algorithm are directly used as the salient places In contrast, the present invention can reduce the misjudgment rate of prominent locations because the feature selection algorithm is used after the clustering algorithm, thereby improving the accuracy of determining prominent locations.

It should be noted that the processor 601 in this embodiment of the present invention may be one processor, or may be a general term for multiple processing elements. For example, the processor 601 may be a central processing unit (Central Processing Unit, referred to as CPU), or a specific integrated circuit (Application Specific Integrated Circuit, referred to as ASIC), or be configured to implement one or more An integrated circuit, for example: one or more microprocessors (digital signal processor, DSP for short), or one or more field programmable gate arrays (Field Programmable Gate Array, FPGA for short).

The memory 602 may be a storage device, or a general term for multiple storage elements, and is used to store executable program codes and the like. And the memory 602 may include random access memory (RAM), and may also include non-volatile memory (non-volatile memory), such as disk memory, flash memory (Flash), and the like.

The bus 603 may be an Industry Standard Architecture (Industry Standard Architecture, ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (Extended Industry Standard Architecture, EISA) bus, etc. The bus 603 can be divided into address bus, data bus, control bus and so on. For ease of representation, only one thick line is used in FIG. 6 , but it does not mean that there is only one bus or one type of bus.

Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be realized by means of software plus necessary general-purpose hardware, and of course also by hardware, but in many cases the former is a better embodiment . Based on this understanding, the essence of the technical solution of the present invention or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product is stored in a readable storage medium, such as a floppy disk of a computer , a hard disk or an optical disk, etc., including several instructions for enabling a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments of the present invention.

The above is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Anyone skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present invention. Should be covered within the protection scope of the present invention.

Claims (6)

1. a kind of method for determining significant place, which is characterized in that the described method includes: being determined according to sample position track significant Place determines attribute and determines that attribute determines the significant place in location track to be processed according to the significant place；

Wherein, described to determine that significant place determines attribute according to sample position track, it specifically includes:

At least one potential significant place on the sample position track is obtained by default clustering algorithm；

The value for obtaining corresponding first attribute set in each potential significant place, obtains attribute information table, described first belongs to Property set includes the conditional attribute and a decision attribute of predetermined number；Wherein, the conditional attribute include speed, acceleration, Record number of the potential significant place in the sample position track, the potential significant place are in the sample position track The number of days of middle appearance, the potential significant place time period of stay and Orientation differences standard deviation；The decision attribute refers to Whether the potential significant place is significant place；

According to the attribute information table, significant place is chosen from first attribute set using default Feature Selection algorithm and is determined Determine attribute；It specifically includes: calculating having for each conditional attribute in first attribute set using default Feature Selection algorithm The value of the P norm of fault tolerance；Maximum value in all values of P norm described in being calculated with fault tolerance is corresponding Conditional attribute be determined as significant place and determine attribute；

Obtain the threshold range that the significant place determines attribute；

It is described to determine that attribute determines the significant place in location track to be processed according to the significant place, it specifically includes:

At least one potential significant place in the location track to be processed is obtained by the default clustering algorithm；

Significant place is determined that potential significant place of the value of attribute in the threshold range is determined as significant place.

2. the method according to claim 1, wherein described calculate described first using default Feature Selection algorithm The value of the P norm with fault tolerance of each conditional attribute in attribute set, specifically includes:

Using the first condition attribute in first attribute set as attribute to be assessed, the set that remaining conditional attribute is constituted It is determined as the second attribute set, the first condition attribute is any conditional attribute；

Game playing algorithm is selected using the user property based on fuzzy entropy, successively judges the attribute to be assessed in second attribute Whether the sharing in game for each destination subset in set wins, and obtains judging result corresponding with each destination subset, institute Stating destination subset includes at least two conditional attributes；

According to all judging results, the value of the P norm with fault tolerance of the attribute to be assessed is calculated.

3. the method according to claim 1, wherein the default clustering algorithm includes distance threshold, points threshold Value and three parameters of time threshold；

It is described that at least one on the sample position track or the location track to be processed is obtained by default clustering algorithm Potential significant place, specifically includes:

Place not labeled in the location track is used as to starting point, the label include be labeled as potential significant place or Noise spot；

Search the target point for being less than or equal to the distance threshold at a distance from the starting point；

If the quantity of the target point found is more than or equal to the points threshold value, and each target point and described The time interval of initial point is all larger than the time threshold, then the starting point and all target points is determined as a cluster, And the starting point is labeled as potential significant place；

After all the points in the cluster are marked, the place that successively marks other not to be labeled.

4. a kind of device for determining significant place, which is characterized in that described device includes:

Processing module, for obtaining at least one potential significant place on sample position track by default clustering algorithm；

Module is obtained, for obtaining the value of corresponding first attribute set in each potential significant place, obtains attribute information Table, first attribute set include the conditional attribute and a decision attribute of predetermined number；Wherein, the conditional attribute includes Record number in the sample position track of speed, acceleration, potential significant place, the potential significant place are described The time period of stay and Orientation differences standard deviation of the number of days, the potential significant place that occur in sample position track；It is described Decision attribute refers to whether the potential significant place is significant place；

The processing module, is also used to according to the attribute information table, using default Feature Selection algorithm from first attribute Significant place is chosen in set determines attribute；

The acquisition module is specifically used for calculating each condition category in first attribute set using default Feature Selection algorithm The value of the P norm with fault tolerance of property；In all values of P norm described in being calculated with fault tolerance most It is worth corresponding conditional attribute greatly and is determined as significant place decision attribute；

The acquisition module is also used to obtain the threshold range that the significant place determines attribute；

The processing module, at least one for being also used to obtain in location track to be processed by the default clustering algorithm are potential Significant place；

The processing module, be also used to determine in the significant place value of attribute in the threshold range it is potential significantly Point is determined as significant place.

5. device according to claim 4, which is characterized in that the acquisition module, also particularly useful for:

Using the first condition attribute in first attribute set as attribute to be assessed, the set that remaining conditional attribute is constituted It is determined as the second attribute set, the first condition attribute is any conditional attribute；

Game playing algorithm is selected using the user property based on fuzzy entropy, successively judges the attribute to be assessed in second attribute Whether the sharing in game for each destination subset in set wins, and obtains judging result corresponding with each destination subset, institute Stating destination subset includes at least two conditional attributes；

According to all judging results, the value of the P norm with fault tolerance of the attribute to be assessed is calculated.

6. device according to claim 4, which is characterized in that the default clustering algorithm includes distance threshold, points threshold Value and three parameters of time threshold；

The processing module, is specifically used for:

Place not labeled in the location track is used as to starting point, the label include be labeled as potential significant place or Noise spot；

Search the target point for being less than or equal to the distance threshold at a distance from the starting point；

If the quantity of the target point found is more than or equal to the points threshold value, and each target point and described The time interval of initial point is all larger than the time threshold, then the starting point and all target points is determined as a cluster, And the starting point is labeled as potential significant place；

After all the points in the cluster are marked, the place that successively marks other not to be labeled.