CN116010830A

CN116010830A - Network taxi driver identification method, device, server and storage medium

Info

Publication number: CN116010830A
Application number: CN202211673152.9A
Authority: CN
Inventors: 孔祥斌; 欧阳秀平; 林敏�; 陈祥; 杨春民; 刘卉芳; 邹俊德; 廖娟; 彭诗雅; 杨沛; 江俊昊; 闫猛; 叶海宁
Original assignee: China United Network Communications Group Co Ltd
Current assignee: China United Network Communications Group Co Ltd
Priority date: 2022-12-26
Filing date: 2022-12-26
Publication date: 2023-04-25

Abstract

The invention provides a network taxi driver identification method, a device, a server and a storage medium, wherein the method comprises the following steps: acquiring a first network address in a driver side App of the network taxi platform and a second network address in a passenger side App, and marking a user accessing the first network address as a network taxi driver if the first and second network addresses do not have the same field; if the same field exists, screening a user set using a network-passing vehicle-closing platform from the user mobile network detail list; counting at least one platform usage feature and at least one spatial movement feature of each user in a user mobile network detail list; and processing the characteristic values of the characteristics, and screening and marking out net bus drivers from all users according to the processed data. The invention realizes the identification of the users with network details but without professional labels by extracting the characteristics from the mobile network details of the users and marking out the network taxi drivers from all platform users according to the characteristics.

Description

Network taxi driver identification method, device, server and storage medium

Technical Field

The present invention relates to the field of wireless communications technologies, and in particular, to a method and apparatus for identifying a driver of a network taxi, a server, and a storage medium.

Background

The user labels of the current operators are mainly related to communication, and the labels related to social attributes such as user occupation and the like are lacking. Community identification of professional labels can be classified into supervised classification, semi-supervised classification, and unsupervised classification according to the degree to which data is truly tagged. Because the general user does not actively disclose professional information to the operator, the user is hardly tagged with the professional tag at all.

In the prior art, a majority of user occupation inference and group discovery recognition are based on more or less labeled datasets, a "fitting" method to known tags. In addition, there are a few techniques that use self-defined rules and thresholds to calculate their confidence in a certain portion of occupation based on some user characteristics, and then use these confidence to generate occupation decisions.

However, the inventors found that the professional inference and group identification methods in the prior art rely on professional labels, selection of professional spaces, rules and reasonability of thresholds, and that it is difficult for users without professional labels to conduct community classification and identification.

Disclosure of Invention

The invention provides a method, a device, a server and a storage medium for identifying a network taxi driver, which are used for solving the problems that the occupation inference and group identification method in the prior art depends on occupation labels, and community division and identification are difficult for users without the occupation labels.

In a first aspect, the present invention provides a method for identifying a driver of a network taxi, including:

acquiring a first network address used for network communication in a driver side App of any of various network vehicle-restraining platforms and a second network address used for network communication in a passenger side App, and judging whether the first network address and the second network address have the same field;

if the same field does not exist in the first network address and the second network address, marking the user accessing the first network address as a network bus driver;

if the same field exists in the first network address and the second network address, screening out users using the network taxi-closing platform in a preset time period from a user mobile network detail list, and constructing a user set of a driver and passenger mixture;

in the user mobile network detail list, counting at least one platform use characteristic of each user in the user set according to a preset acquisition interval;

in the user mobile network detail list, counting at least one space movement characteristic of each user in the user set according to a preset acquisition interval;

normalizing the characteristic values of the platform use characteristic and the space movement characteristic, and stacking all the normalized characteristic values to obtain a behavior characteristic vector of each user;

Clustering is carried out on the behavior feature vectors of all users to obtain a target clustering result;

and screening out users of the net taxi drivers from all the users according to the target clustering result, and marking the net taxi drivers as net taxi drivers.

In one possible design, the platform usage feature includes two: the sum of the longest continuous access network taxi-taking platform duration and the continuous access network taxi-taking platform duration; correspondingly, in the user mobile network detail list, counting at least one platform use characteristic of each user in the user set according to a preset collection interval, including: screening out network addresses belonging to a network taxi platform in the mobile network detail list of the current user, arranging the starting time of the network addresses belonging to the network taxi platform according to ascending order, wherein the number of entries of the network addresses belonging to the network taxi platform is n; initializing a list L= [0] for recording the duration of the nth network address continuous access network vehicle platform, and initializing an initial value of a memory variable min_last, wherein the value of the min_last is the minutes of the recorded start time of the 1 st network address; calculating the difference value between the minutes of the start time of the recorded a-th network address and the min_last, and if the difference value is 0, ignoring the record; let L [ -1] =l [ -1] +1, where L [ -1] represents the value of the last bit in L, if the difference is 1; if the difference value is greater than 1, inserting 0 into the last bit in L, and reassigning the min_last to be the minute of the start time of the a-th record; judging whether a is less than or equal to n; if yes, let a=a+1, and repeat the step of calculating the difference between the minute at which the start time of the a-th record is located and the min_last; if not, taking the maximum value max (L) in L and the sum (L) of each numerical value in L as the platform use characteristics of the current user on the same day; where max (L) represents the length of time for which the longest continuous access network about the vehicle platform, and sum (L) represents the sum of the lengths of time for which the continuous access network about the vehicle platform.

In one possible design, the spatial movement feature includes two: the number of LACs connected and the Jaccard correlation coefficient of the LAC set; correspondingly, in the user mobile network detail list, counting at least one space movement characteristic of each user in the user set according to a preset collection interval, and comprising the following steps: removing duplication of all the connected location area codes LAC of the current user on the a-th day, constructing a set, marking the set as LACs_last, and recording the number of elements in the set LACs_last; duplicate removal is carried out on all the connected location area codes LAC of the current user at the (a+1) th day, a set is constructed, the set is recorded as LACs_current, and the number of elements in the set LACs_current is recorded; dividing the number of elements in the intersection of the set LACs_current and the set LACs_last by the number of elements in the union of the set LACs_current and the set LACs_last to obtain a Jaccard correlation coefficient; and stacking the set size of the location area codes LAC connected by the user and Jaccard correlation coefficients to obtain the space movement characteristic, wherein the set size of the location area codes LAC connected by the user represents the number of the connected LACs, and the Jaccard correlation coefficients represent the Jaccard correlation coefficients of the LAC set.

In one possible design, the normalizing the feature values of the platform usage feature and the spatial movement feature stacks all the normalized feature values to obtain a behavior feature vector of each user, including: respectively calculating the average value and variance of the platform use characteristics and the space movement characteristics of each user in all users in a preset number of days; dividing the difference value between the characteristic values of the platform use characteristics and the space movement characteristics of each user in a preset day and the average value by the variance to obtain the characteristic values of the standardized platform use characteristics and the standardized space movement characteristics; and splicing the characteristic values of the standardized platform use characteristic and the space movement characteristic into a vector to obtain a behavior characteristic vector of each user.

In one possible design, the clustering processing is performed on the behavior feature vectors of all the users to obtain a target clustering result, including: k-means clustering is carried out on the feature vectors by adopting different clustering numbers to obtain a plurality of clustering results, and contour coefficients of the clustering results are calculated; and screening target clustering results with the contour coefficient larger than a preset value from all the clustering results to obtain the target clustering number corresponding to the target clustering results.

In one possible design, the method for screening out users of the network about car drivers according to the target clustering result, and marking the users as the network about car drivers includes: the classes in the clusters corresponding to the target cluster number are arranged in descending order according to the characteristic values, the classes in the clusters are divided into two groups according to the arrangement order, the first group is assumed to be a net car driver, and the second group is assumed to be a passenger; calculating the user quantity ratio in the first group and the second group, screening out two groups with the smallest absolute value of the difference value between the user quantity ratio and the actual ratio, marking the users in the first group of the two screened groups as network taxi drivers, and marking the users in the second group as passengers; the actual ratio is the driver-passenger ratio of the network about car industry which is obtained through pre-investigation.

In a second aspect, the present invention provides a net restraint vehicle driver identification device, comprising:

the acquisition module is used for acquiring a first network address used for network communication in a driver side App of any of various network vehicle-restraining platforms and a second network address used for network communication in a passenger side App and judging whether the first network address and the second network address have the same field or not;

The first marking module is used for marking the user accessing the first network address as a network bus driver if the same field does not exist in the first network address and the second network address;

the construction module is used for screening out users using the network taxi-taking platform in a preset time period from the user mobile network detail list if the same fields exist in the first network address and the second network address, and constructing a user set of a driver and passenger mixture;

the first statistics module is used for counting at least one platform use characteristic of each user in the user set according to a preset acquisition interval in the user mobile network detail list;

the second statistics module is used for counting at least one space movement characteristic of each user in the user set according to a preset acquisition interval in the user mobile network detail list;

the standardized module is used for standardizing the characteristic values of the platform use characteristic and the space movement characteristic, and stacking all the standardized characteristic values to obtain a behavior characteristic vector of each user;

the clustering module is used for carrying out clustering processing on the behavior feature vectors of all users to obtain a target clustering result;

And the second marking module is used for screening out users of the network taxi drivers from all the users according to the target clustering result and marking the users as the network taxi drivers.

In one possible design, the first statistical module is specifically configured to, among other things, the platform usage characteristics include two: the sum of the longest continuous access network taxi-taking platform duration and the continuous access network taxi-taking platform duration; screening out network addresses belonging to a network taxi platform in the mobile network detail list of the current user, arranging the starting time of the network addresses belonging to the network taxi platform according to ascending order, wherein the number of entries of the network addresses belonging to the network taxi platform is n; initializing a list L= [0] for recording, and initializing an initial value of a memory variable min_last, wherein the value of the min_last is the minute at which the start time of the 1 st record is located; calculating the difference value between the minute at which the start time of the a record is located and the min_last, and if the difference value is 0, neglecting the record; let L [ -1] =l [ -1] +1, where L [ -1] represents the value of the last bit in L, if the difference is 1; if the difference value is greater than 1, inserting 0 into the last bit in L, and reassigning the min_last to be the minute of the start time of the a-th record; judging whether a is less than or equal to n; if yes, let a=a+1, and repeat the step of calculating the difference between the minute at which the start time of the a-th record is located and the min_last; if not, taking the maximum value max (L) in L and the sum (L) of each numerical value in L as the platform use characteristics of the current user on the same day; where max (L) represents the length of time for which the longest continuous access network about the vehicle platform, and sum (L) represents the sum of the lengths of time for which the continuous access network about the vehicle platform.

In a third aspect, the present invention provides a server comprising: at least one processor and memory;

the memory stores computer-executable instructions;

the at least one processor executes the computer-executable instructions stored by the memory such that the at least one processor performs the above described first aspect and the various possible designs of the first aspect.

In a fourth aspect, the present invention provides a computer storage medium having stored therein computer executable instructions which when executed by a processor implement the above described first aspect and the various possible designs of the first aspect.

According to the network taxi driver identification method, device, server and storage medium, platform use features and space movement features of the user are extracted from the mobile network detail list of the user, and the features are processed and clustered, so that the network taxi drivers are marked from all platform users, professional community division and identification can be carried out on users with network detail list and no professional label, and the user does not need to rely on a threshold value or rule set by experience.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the drawings that are needed in the embodiments or the description of the prior art will be briefly described below, it will be obvious that the drawings in the following description are some embodiments of the present invention, and that other drawings can be obtained according to these drawings without inventive effort to a person skilled in the art.

Fig. 1 is a schematic flow chart of a method for identifying a driver of a network bus according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of a second method for identifying a driver of a network bus according to an embodiment of the present invention;

FIG. 3 is a second flowchart of a method for identifying a driver of a network bus according to an embodiment of the present invention;

fig. 4 is a schematic structural diagram of a network bus driver recognition device according to an embodiment of the present invention;

fig. 5 is a schematic hardware structure of a server according to an embodiment of the present invention.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

It should be noted that, the user information (including but not limited to user equipment information, user personal information, etc.) and the data (including but not limited to data for analysis, stored data, presented data, etc.) related to the present application are information and data authorized by the user or fully authorized by each party, and the collection, use and processing of the related data need to comply with the related laws and regulations and standards of the related country and region, and provide corresponding operation entries for the user to select authorization or rejection.

The label of the specific occupation is the basis of the enterprise for personalized product research and development and differentiated accurate marketing. The user labels of the current operators are mainly related to communication, and the labels related to social attributes such as user occupation and the like are lacking. Community identification of professional labels is generally categorized as a classification problem in machine learning. Classification problems can be classified into supervised classification, semi-supervised classification, and unsupervised classification depending on the degree to which data is truly tagged. Because the general user does not actively disclose professional information to the operator, the user is hardly tagged with the professional tag at all. In the prior art, a majority of user occupation inference and group discovery identification are based on more or less marked data sets, and this type of technique finds an operation from a number of features of a person known to be engaged in an occupation or known not to be engaged in an occupation, and maps each feature to a probability that he is engaged in the occupation, so that the error rate of inference using the probability is the lowest, which is a "fitting" method to known labels. In addition, there are a few techniques that use self-defined rules and thresholds to calculate their confidence in a certain portion of occupation based on some user characteristics, and then use these confidence to generate occupation decisions. However, the inventor found that the professional inference and group identification methods in the prior art depend on the reasonability of professional labels, selection of professional spaces, rules and thresholds, it is difficult for users without professional labels to conduct community division and identification, and it is difficult to dynamically adjust the criteria of professional inference and group identification as objective conditions change.

In order to solve the technical problems, the embodiment of the invention provides the following technical scheme: the network booking car drivers are marked from all platform users by extracting platform use characteristics and space movement characteristics of the users from the mobile network details of the users and processing and clustering the characteristics.

Fig. 1 is a schematic flow chart of a method for identifying a driver of a network bus in an embodiment of the present invention, where the execution subject of the embodiment may be a server or other computer devices, and the embodiment is not limited herein. As shown in fig. 1, the method includes:

s101: and acquiring a first network address used for network communication in a driver side App of any of various network vehicle-restraining platforms and a second network address used for network communication in a passenger side App.

In this embodiment, the first network address and the second network address refer to a uniform resource locator (Uniform Resource Locator, URL). The method of acquiring the first network address and the second network address may adopt a method of network inquiry or packet grasping.

S102: judging whether the same field exists in the first network address and the second network address, if not, executing step S103, and if so, executing step S104.

In this embodiment, the judgment whether the same field exists in the first network address and the second network address is as follows: the URL used by the client App of the drip platform may include an "xiaojukeji" field, or a "di" field. If the URLs used by the driver side App and the passenger side App do not have the same field, step S103 is performed. If the URLs used by the driver side App and the passenger side App are relatively similar, and the driver or passenger identity of the user cannot be distinguished only by the URL, step S104 must be performed.

S103: the user who has accessed the first network address is marked as a netbook vehicle driver.

In this embodiment, because the driver side App performs network communication at the first network address, the user who accesses the first network address may be directly marked as the network bus driver.

S104: and screening users using the network-passing vehicle-restraining platform in a preset time period from the user mobile network detail list, and constructing a user set of a driver and passenger mixture.

In this embodiment, the preset time period may be set according to actual situations, and may be set to one week or two weeks; for example, if a vehicle driver who has not been able to calculate a net bus for a week is considered to be not going out, the preset time period may be set to one week, and if the tolerance of this tag is considered to be greater, the preset time period may be set to two weeks.

S105: in the user mobile network detail list, at least one platform use characteristic of each user in the user set is counted according to a preset collection interval.

In this embodiment, the platform usage features include two: the sum of the longest continuous access network taxi-stand time length and the continuous access network taxi-stand time length. The reason why the sum of the time length of the longest continuous access network taxi-offered platform and the time length of the continuous access network taxi-offered platform is selected as the platform usage feature is as follows: the driver needs to communicate frequently with the network about car platform server through the mobile network in order to serve the passengers, and also needs to communicate with the server for a long time in order to maintain the service or listening state, the passengers do not have such great demands in both aspects, so the main difference between the driver and the passengers in the access behavior of the network about car platform is that the time continuity and the overall strength of the network about car App are used, so the longest daily continuous access time length of the network about car platform can be used as a description of the continuity, and the sum of the daily continuous access time lengths of the network about car platform can also be used as a description of the overall use strength.

The information of the user mobile network detailed list contains important information such as time, connected position area code, accessed network address and the like, and the behavior of the user can be reflected to a great extent. Because all important nodes and operations need to be controlled by an online platform in the process of providing service for passengers by a network taxi driver and frequently interact with a server of the network taxi platform through a mobile network, in user data mastered by an operator, the mobile network detail list is an ideal data source for carrying out behavior modes of the network taxi driver and the passengers; therefore, the embodiment of the invention utilizes different behavior patterns of the driver and the passenger to conduct community identification of the network appointment vehicle driver.

S106: in the user mobile network detail list, at least one space movement characteristic of each user in the user set is counted according to a preset acquisition interval.

In this embodiment, the spatial movement features include two: the number of concatenated location area codes (location area code, LAC) and Jaccard correlation coefficients of the LAC set. The number of connected LACs and Jaccard correlation coefficients of the LAC set are selected as the spatial movement characteristics for the following reasons: the driver often drives own vehicles in urban space in the operation process for half a day or all the day, or moves at the driving speed all the time for taking the passengers or going to the places with multiple orders, so that the total mileage of the driver in one day is longer than that of the passengers; in addition, the activity of the driver is subjected to random order generation, and the random order generation has strong randomness, so that the track of the driver is more random than that of a common passenger, which is determined by a regular life mode. It is believed that the main difference in the spatial movement behavior of the driver and passenger is mileage, which can be characterized by the number of passing LACs, and randomness, which can be characterized by a correlation coefficient with the trajectory of the previous day.

S107: and normalizing the characteristic values of the platform using characteristic and the space moving characteristic, and stacking all the normalized characteristic values to obtain the behavior characteristic vector of each user.

In this embodiment, the average and variance of the platform usage feature and the spatial movement feature of each of all the users in a preset number of days are calculated respectively. Dividing the difference between the characteristic value of the platform use characteristic and the space movement characteristic of each user in a preset day and the average value by the variance to obtain the characteristic value of the standardized platform use characteristic and the standardized space movement characteristic. And splicing the characteristic values of the standardized platform use characteristic and the space movement characteristic into a vector to obtain the behavior characteristic vector of each user.

The known platform usage features in steps S105 and S106 include two: the sum of the longest continuous access network taxi-stand time length and the continuous access network taxi-stand time length. The spatial movement features include two: the number of LACs connected and the Jaccard correlation coefficient of the LAC set.

Specifically, if the preset number of days is m days, the features of each user may be summarized as follows: the sum of the durations of the m longest continuous access network provisioning platforms, the number of m connected LACs, and the Jaccard correlation coefficient for the m LAC sets.

Illustratively, taking the duration of m longest continuous access network taxi-taking platforms as an example, the step of normalizing the feature is:

The average mean and variance std of the duration of m longest continuous access network vehicle-restraining platforms are calculated, and the following formula is used for carrying out standardized operation:

wherein x' is the standardized m longest duration of continuous access network vehicle-restraining platforms; x is the duration of the m longest continuous access network taxi-taking platforms, mean is the average of the duration of the m longest continuous access network taxi-taking platforms, and std is the variance of the duration of the m longest continuous access network taxi-taking platforms.

Similarly, the step of normalizing the three features, that is, the sum of the durations of the m continuous access network taxi-taking platforms, the number of m connected LACs and the Jaccard correlation coefficient of the m LAC sets, is the same as the step of normalizing the durations of the m longest continuous access network taxi-taking platforms, so that the description thereof is omitted here.

S108: and clustering the behavior feature vectors of all the users to obtain a target clustering result.

In the embodiment, K-means clustering is carried out on the feature vectors by adopting different clustering numbers to obtain a plurality of clustering results, and the contour coefficients of the clustering results are calculated; and screening target clustering results with contour coefficients larger than a preset value from all the clustering results to obtain target clustering numbers corresponding to the target clustering results.

Specifically, 2, 3, 4 and 5 are adopted as the clustering number to perform K-means clustering on the feature vectors respectively, a plurality of clustering results are obtained, the contour coefficient is calculated for each clustering result respectively, and the clustering result with the contour coefficient larger than 0.6 is selected. The contour coefficient is an index for evaluating the clustering effect and is used for describing the contour definition of each class after clustering. The range of the values of the contour coefficients is [ -1,1], and the larger the contour coefficients are, the better the clustering effect is.

For example, when the number of clusters is 2, the profile factor of the clustering result is about 0.9; when the number of clusters is 3, the contour coefficient of the clustering result is about 0.7; when the number of clusters is 4 or higher, the profile factor of the clustering result is about 0.2. In such a case, the method of taking the number of clusters of 4 or higher should be abandoned, and the clustering results of the numbers of clusters of 2 and 3 should be studied with emphasis. Namely, in this embodiment, the number of target clusters corresponding to the target cluster result is 2 and 3.

S109: and screening out users of the network taxi drivers from all the users according to the target clustering result, and marking the users as the network taxi drivers.

In this embodiment, the classes in the clusters corresponding to the number of target clusters are arranged in descending order according to the feature values, and the classes in the clusters are divided into two groups according to the arrangement order, wherein the first group is assumed to be a network taxi driver, and the second group is assumed to be a passenger. Calculating the user quantity ratio in the first group and the second group, screening out two groups with the smallest absolute value of the difference value between the user quantity ratio and the actual ratio, marking the users in the first group of the two screened groups as network bus drivers, and marking the users in the second group as passengers; the actual ratio is the driver-passenger ratio of the network about car industry which is obtained through pre-investigation.

Specifically, given that the number of target clusters is known to be 2 and 3 in step 208, for a cluster with a number of clusters of 2, if the resulting 2 classes (0, 1) are arranged to be 1, 0 in descending order of the eigenvalues, then the possible groupings are only [ [1], [0] ]. For a cluster with a cluster number of 3, assuming that the resulting 3 classes (0, 1, 2) are arranged as 1,2, 0 in descending order of eigenvalue, the possible groupings are [ [1], [2,0] ] and [ [1,2], [0] ]. For a total of 4 groups described above, the first group is assumed to be the net bus driver and the second group is assumed to be the passenger, respectively. And calculating the user quantity ratio of the first group and the second group in each group to obtain the user quantity ratio of each group.

The ratio of the driver to the passenger in the network vehicle industry obtained through pre-investigation is recorded as the actual ratio. And comparing the user quantity ratios of the groups with the actual ratio, screening out the user quantity ratio with the minimum absolute value of the difference value between the user quantity ratio and the actual ratio, marking the first group of users in the group corresponding to the screened user quantity ratio as a network bus driver, and marking the second group of users as passengers. After the marking result of the net bus driver is obtained, the marking result is sent to a display terminal or a display large screen for display. In summary, according to the network taxi driver identification method provided by the embodiment, platform use features and space movement features of the user are extracted from the mobile network detail list of the user, and the features are processed and clustered, so that the network taxi drivers are marked from all platform users, professional community division and identification can be performed on users with network detail list and no professional label, and the threshold or rule set by human experience is not required.

Fig. 2 is a schematic diagram of a second method for identifying a driver of a network bus according to an embodiment of the present invention. In the embodiment of the present invention, on the basis of the embodiment provided in fig. 1, a specific implementation method of the usage feature of at least one platform for counting each user in the user set in step S105 is described in detail. As shown in fig. 2, the method includes:

s201: in the mobile network detail list of the user, the network address belonging to the network about car platform in the mobile network detail list of the current user is screened, the starting time of the network address belonging to the network about car platform is arranged according to ascending order, and the number of entries of the network address belonging to the network about car platform is n.

In this embodiment, in the user mobile network detail list, the network address belonging to the network about platform in the mobile network detail list of the current user is selected according to the field in the network address. For example, the network address contains a "di" field, and the network address is considered to belong to the network about vehicle platform. And arranging the starting time of the n screened network addresses belonging to the network about vehicle platform according to an ascending order.

S202: initializing a list L= [0] for recording the duration of the continuous access of the nth network address to the network about car platform, and initializing an initial value of a memory variable min_last.

In this embodiment, l= [0] represents a duration of continuous access to the network vehicle platform by the network address of the first bit in L. The value of min_last is the minutes at which the recorded start time of the 1 st network address is located.

S203: calculating the difference value between the minutes of the start time of the recorded a-th network address and the min_last, and if the difference value is 0, ignoring the record; if the difference is 1, step S204 is performed; if the difference is greater than 1, step S205 is performed.

In this embodiment, the difference between the minute at which the recorded a-th network address starts and the minute_last (i.e. the minute at which the recorded 1-th network address starts) is taken, and if the difference is 0, it is indicated that the a-th network address is the 1 st network address, so the record is ignored.

S204: let L [ -1] =l [ -1] +1.

In this embodiment, L < -1 > represents the last value in L, i.e. the duration of the continuous access network about the platform of the last network address of the record access.

S205: the last bit in L is inserted with 0 while min_last is reassigned to the minute at which the start time of the recorded a-th network address is located.

In this embodiment, if the difference between the minute at which the recorded a-th network address starts and the minute at which the recorded 1-th network address starts is greater than 1, the min_last is reassigned to the minute at which the recorded a-th network address starts. The last bit of 0 in L represents that the continuous access of the previous section of network about car platform is recorded, a new section of continuous access needs to be recorded, the continuous use time length of the section is initialized to 0, and the L constructed by the step can record the time length of each section of continuous access of the network about car platform so as to take out the maximum value later.

S206: judging whether a is less than or equal to n, if so, making a=a+1 and re-executing step S203; if not, go to step S207.

In this embodiment, if a is less than or equal to n, it is indicated that the platform usage feature in all n network addresses belonging to the network about vehicle platform has not been acquired yet, and a=a+1 needs to be continuously acquired. If a is greater than n, the platform usage characteristics of all n network addresses belonging to the network about vehicle platform are acquired.

S207: taking the maximum value max (L) in L and the sum (L) of each numerical value in L as the platform use characteristics of the current user on the same day;

in this embodiment, max (L) represents the length of time for which the network about the platform is continuously accessed, and sum (L) represents the sum of the lengths of time for which the network about the platform is continuously accessed.

In summary, according to the network taxi driver identification method provided by the embodiment, the starting time and the using time of the network address belonging to the network taxi platform are calculated, the sum of the longest continuous access time length of the network taxi platform and the continuous access time length of the network taxi platform is selected from the user mobile network detail list and is used as the platform using characteristic, the threshold value or rule set by relying on experience is not manually used, and the identification of the network taxi driver is possible according to different behavior modes of the driver and the passenger.

Fig. 3 is a schematic diagram of a second method for identifying a driver of a network bus according to an embodiment of the present invention. In the embodiment of the present invention, a detailed description is given of a specific implementation method of at least one spatial movement feature of each user in the statistics user set in step S206 based on the embodiment provided in fig. 1. As shown in fig. 3, the method includes:

s301: and (3) de-duplicating all the connected location area codes LAC on the current user's a day, constructing a set, recording the set as LACs_last, and recording the number of elements in the set LACs_last.

In this embodiment, the location area code LAC to which the user is connected may be used to represent the mileage of the user space movement. The position area codes connected by the user are subjected to de-duplication processing, so that statistics on the space movement of the user can be more accurate, and repeated statistics is avoided.

S302: and (3) de-duplicating all the connected location area codes LAC on the (a+1) th day of the current user, constructing a set, recording the set as LACs_current, and recording the number of elements in the set LACs_current.

The method for constructing the set in this embodiment is the same as that in step S301, and will not be described here again.

S303: dividing the number of elements in the intersection of the set LACs_current and the set LACs_last by the number of elements in the union of the set LACs_current and the set LACs_last to obtain a Jaccard correlation coefficient.

In this embodiment, the calculation method of the Jaccard correlation coefficient is as follows:

wherein, |·| represents the number of elements in the collection; LACs_Current n LACs_last represents the intersection of the set LACs_Current and the set LACs_last; lacs_current_last represents the union of the set lacs_current and the set lacs_last.

S304: and stacking the set size of the location area codes LAC connected by the user and the Jaccard correlation coefficient to obtain the space movement characteristic.

In this embodiment, the set size of the location area codes LAC connected by the user represents the number of LACs connected, and the Jaccard correlation coefficient represents the Jaccard correlation coefficient of the LAC set.

In summary, according to the method for identifying the internet-of-vehicle driver provided by the embodiment, the number of the LACs connected to the internet-of-vehicle platform and the Jaccard correlation coefficient of the LAC set are calculated, the number of the connected LACs and the Jaccard correlation coefficient of the LAC set are screened out from the user mobile network detail list to serve as space movement characteristics, the threshold value or rule set by experience is not manually relied on, and the identification of the internet-of-vehicle driver is possible according to different behavior modes of the driver and the passenger.

Fig. 4 is a schematic structural diagram of a network bus driver recognition device according to an embodiment of the present invention. As shown in fig. 4, the net restraint vehicle driver recognition apparatus includes: an acquisition module 401, a first labeling module 402, a construction module 403, a first statistics module 404, a second statistics module 405, a normalization module 406, a clustering module 407, and a second labeling module 408.

The obtaining module 401 is configured to obtain a first network address used for network communication in a driver end App of any of multiple network vehicle-restraining platforms and a second network address used for network communication in a passenger end App, and determine whether the same field exists in the first network address and the second network address.

The first marking module 402 is configured to mark the user who accesses the first network address as a network bus driver if the same field does not exist.

And the construction module 403 is configured to screen out the users using the network-passing vehicle-restraining platform in a preset time period from the user mobile network details if the same field exists, and construct a user set of a driver and passenger mixture.

A first statistics module 404, configured to count, in the user mobile network details, at least one platform usage feature of each user in the user set according to a preset collection interval.

A second statistics module 405, configured to count at least one spatial movement characteristic of each user in the user set according to a preset collection interval in the user mobile network detail list.

And the normalization module 406 is configured to normalize feature values of the platform usage feature and the spatial movement feature, and stack all the normalized feature values to obtain a behavior feature vector of each user.

And the clustering module 407 is used for clustering the behavior feature vectors of all the users to obtain a target clustering result.

And the second marking module 408 is used for screening out users of the network taxi drivers from all the users according to the target clustering result and marking the users as the network taxi drivers.

In one possible implementation, the first statistics module 404 is specifically configured to, among other things, include two platform usage characteristics: the sum of the longest continuous access network taxi-taking platform duration and the continuous access network taxi-taking platform duration; screening out network addresses belonging to the network about vehicle platform in the mobile network detail list of the current user, arranging the starting time of the network addresses belonging to the network about vehicle platform according to ascending order, wherein the number of entries of the network addresses belonging to the network about vehicle platform is n; initializing a list L= [0] for recording, initializing an initial value of a memory variable min_last, wherein the value of the min_last is the minute at which the start time of the 1 st record is located; calculating the difference value between the minute at which the start time of the a record is located and the min_last, and if the difference value is 0, ignoring the record; let L [ -1] =l [ -1] +1, where L [ -1] represents the value of the last bit in L, if the difference is 1; if the difference is greater than 1, inserting 0 into the last bit in L, and reassigning min_last to be the minute of the start time of the a record; judging whether a is less than or equal to n; if yes, let a=a+1, and repeat the step of calculating the difference between the minutes at which the start time of the a-th record is located and the min_last; if not, taking the maximum value max (L) in L and the sum (L) of each numerical value in L as the platform use characteristics of the current user on the same day; where max (L) represents the length of time for which the longest continuous access network about the vehicle platform, and sum (L) represents the sum of the lengths of time for which the continuous access network about the vehicle platform.

In one possible implementation, the second statistics module 405 is specifically configured to, among other things, include two spatial movement features: the number of LACs connected and the Jaccard correlation coefficient of the LAC set; removing duplication of all the connected location area codes LAC of the current user on the a-th day, constructing a set, marking the set as LACs_last, and recording the number of elements in the set LACs_last; duplicate removal is carried out on all the connected location area codes LAC of the current user at the (a+1) th day, a set is constructed, the set is recorded as LACs_current, and the number of elements in the set LACs_current is recorded; dividing the number of elements in the intersection of the set LACs_current and the set LACs_last by the number of elements in the union of the set LACs_current and the set LACs_last to obtain a Jaccard correlation coefficient; and stacking the set size of the location area codes LAC connected by the user and Jaccard correlation coefficients to obtain a space movement characteristic, wherein the set size of the location area codes LAC connected by the user represents the number of connected LACs, and the Jaccard correlation coefficients represent Jaccard correlation coefficients of the LAC set.

In one possible implementation, the normalization module 406 is specifically configured to calculate an average and variance of the platform usage feature and the spatial movement feature of each of all the users in a preset number of days, respectively; dividing the difference between the characteristic value of the platform use characteristic and the space movement characteristic of each user in a preset day and the average value by the variance to obtain the characteristic value of the standardized platform use characteristic and the standardized space movement characteristic; and splicing the characteristic values of the standardized platform use characteristic and the space movement characteristic into a vector to obtain the behavior characteristic vector of each user.

In one possible implementation manner, the clustering module 407 is specifically configured to perform K-means clustering on the feature vectors by using different numbers of clusters to obtain a plurality of clustering results, and calculate a profile coefficient of each clustering result; and screening target clustering results with contour coefficients larger than a preset value from all the clustering results to obtain target clustering numbers corresponding to the target clustering results.

In one possible implementation manner, the second marking module 408 is specifically configured to arrange the classes in the clusters corresponding to the target cluster number in descending order according to the feature values, and divide the classes in the clusters into two groups according to the arrangement order, where the first group is assumed to be a network bus driver, and the second group is assumed to be a passenger; calculating the user quantity ratio in the first group and the second group, screening out two groups with the smallest absolute value of the difference value between the user quantity ratio and the actual ratio, marking the users in the first group of the two screened groups as network bus drivers, and marking the users in the second group as passengers; the actual ratio is the driver-passenger ratio of the network about car industry which is obtained through pre-investigation.

The device provided in this embodiment may be used to implement the technical solution of the foregoing method embodiment, and its implementation principle and technical effects are similar, and this embodiment will not be described herein again.

Fig. 5 is a schematic hardware structure of a server according to an embodiment of the present invention. As shown in fig. 5, the server 50 of the present embodiment includes: a processor 501 and a memory 502; wherein the method comprises the steps of

A memory 502 for storing computer-executable instructions;

the processor 501 is configured to execute computer-executable instructions stored in the memory to implement the steps executed by the server in the above embodiment. Reference may be made in particular to the relevant description of the embodiments of the method described above.

Alternatively, the memory 502 may be separate or integrated with the processor 501.

When the memory 502 is provided separately, the server further comprises a bus 503 for connecting the memory 502 and the processor 501.

The embodiment of the invention also provides a computer storage medium, wherein computer execution instructions are stored in the computer storage medium, and when a processor executes the computer execution instructions, the method for identifying the network bus driver is realized.

The embodiment of the invention also provides a computer program product, which comprises a computer program, and when the computer program is executed by a processor, the method for identifying the internet-based vehicle driver is realized. The embodiment of the invention also provides a computer program product, which comprises a computer program, and when the computer program is executed by a processor, the method for identifying the internet taxi driver is realized.

In the several embodiments provided by the present invention, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described device embodiments are merely illustrative, e.g., the division of modules is merely a logical function division, and there may be additional divisions of actual implementation, e.g., multiple modules may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or modules, which may be in electrical, mechanical, or other forms.

The modules illustrated as separate components may or may not be physically separate, and components shown as modules may or may not be physical units, may be located in one place, or may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to implement the solution of this embodiment.

In addition, each functional module in the embodiments of the present invention may be integrated in one processing unit, or each module may exist alone physically, or two or more modules may be integrated in one unit. The units formed by the modules can be realized in a form of hardware or a form of hardware and software functional units.

The integrated modules, which are implemented in the form of software functional modules, may be stored in a computer readable storage medium. The software functional modules described above are stored in a storage medium and include instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) or processor to perform some steps of the methods of the various embodiments of the present application.

It should be understood that the above processor may be a central processing unit (Central Processing Unit, abbreviated as CPU), but may also be other general purpose processors, digital signal processors (Digital Signal Processor, abbreviated as DSP), application specific integrated circuits (Application Specific Integrated Circuit, abbreviated as ASIC), etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of a method disclosed in connection with the present invention may be embodied directly in a hardware processor for execution, or in a combination of hardware and software modules in a processor for execution.

The memory may comprise a high-speed RAM memory, and may further comprise a non-volatile memory NVM, such as at least one magnetic disk memory, and may also be a U-disk, a removable hard disk, a read-only memory, a magnetic disk or optical disk, etc.

The bus may be an industry standard architecture (Industry Standard Architecture, ISA) bus, an external device interconnect (Peripheral Component Interconnect, PCI) bus, or an extended industry standard architecture (Extended Industry Standard Architecture, EISA) bus, among others. The buses may be divided into address buses, data buses, control buses, etc. For ease of illustration, the buses in the drawings of the present application are not limited to only one bus or one type of bus.

The storage medium may be implemented by any type or combination of volatile or nonvolatile memory devices such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk. A storage media may be any available media that can be accessed by a general purpose or special purpose computer.

An exemplary storage medium is coupled to the processor such the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium may be integral to the processor. The processor and the storage medium may reside in an application specific integrated circuit (Application Specific Integrated Circuits, ASIC for short). It is also possible that the processor and the storage medium reside as discrete components in an electronic device or a master device.

Those of ordinary skill in the art will appreciate that: all or part of the steps for implementing the method embodiments described above may be performed by hardware associated with program instructions. The foregoing program may be stored in a computer readable storage medium. The program, when executed, performs steps including the method embodiments described above; and the aforementioned storage medium includes: various media that can store program code, such as ROM, RAM, magnetic or optical disks.

Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.

Claims

1. The network vehicle driver identification method is characterized by being applied to a server and comprising the following steps of:

2. The method of claim 1, wherein the platform usage feature comprises two: the sum of the longest continuous access network taxi-taking platform duration and the continuous access network taxi-taking platform duration;

correspondingly, in the user mobile network detail list, counting at least one platform use characteristic of each user in the user set according to a preset collection interval, including:

screening out network addresses belonging to a network taxi platform in the mobile network detail list of the current user, arranging the starting time of the network addresses belonging to the network taxi platform according to ascending order, wherein the number of entries of the network addresses belonging to the network taxi platform is n;

initializing a list L= [0] for recording the duration of the nth network address continuous access network vehicle platform, and initializing an initial value of a memory variable min_last, wherein the value of the min_last is the minutes of the recorded start time of the 1 st network address;

calculating the difference value between the minutes of the start time of the recorded a-th network address and the min_last, and if the difference value is 0, ignoring the record;

let L [ -1] =l [ -1] +1, where L [ -1] represents the value of the last bit in L, if the difference is 1;

If the difference value is greater than 1, inserting 0 into the last bit in L, and reassigning the min_last to be the minute of the start time of the a-th record;

judging whether a is less than or equal to n;

if yes, let a=a+1, and repeat the step of calculating the difference between the minute at which the start time of the a-th record is located and the min_last; if not, taking the maximum value max (L) in L and the sum (L) of each numerical value in L as the platform use characteristics of the current user on the same day; where max (L) represents the length of time for which the longest continuous access network about the vehicle platform, and sum (L) represents the sum of the lengths of time for which the continuous access network about the vehicle platform.

3. The method of claim 1, wherein the spatial movement feature comprises two: the number of LACs connected and the Jaccard correlation coefficient of the LAC set;

correspondingly, in the user mobile network detail list, counting at least one space movement characteristic of each user in the user set according to a preset collection interval, and comprising the following steps:

removing duplication of all the connected location area codes LAC of the current user on the a-th day, constructing a set, marking the set as LACs_last, and recording the number of elements in the set LACs_last;

duplicate removal is carried out on all the connected location area codes LAC of the current user at the (a+1) th day, a set is constructed, the set is recorded as LACs_current, and the number of elements in the set LACs_current is recorded;

Dividing the number of elements in the intersection of the set LACs_current and the set LACs_last by the number of elements in the union of the set LACs_current and the set LACs_last to obtain a Jaccard correlation coefficient;

and stacking the set size of the location area codes LAC connected by the user and Jaccard correlation coefficients to obtain the space movement characteristic, wherein the set size of the location area codes LAC connected by the user represents the number of the connected LACs, and the Jaccard correlation coefficients represent the Jaccard correlation coefficients of the LAC set.

4. The method according to claim 1, wherein normalizing the feature values of the platform usage feature and the spatial movement feature, stacking all normalized feature values to obtain a behavior feature vector of each user, comprises:

respectively calculating the average value and variance of the platform use characteristics and the space movement characteristics of each user in all users in a preset number of days;

dividing the difference value between the characteristic values of the platform use characteristics and the space movement characteristics of each user in a preset day and the average value by the variance to obtain the characteristic values of the standardized platform use characteristics and the standardized space movement characteristics;

And splicing the characteristic values of the standardized platform use characteristic and the space movement characteristic into a vector to obtain a behavior characteristic vector of each user.

5. The method of claim 1, wherein the clustering the behavior feature vectors of all users to obtain a target clustering result comprises:

k-means clustering is carried out on the feature vectors by adopting different clustering numbers to obtain a plurality of clustering results, and contour coefficients of the clustering results are calculated;

and screening target clustering results with the contour coefficient larger than a preset value from all the clustering results to obtain the target clustering number corresponding to the target clustering results.

6. The method according to any one of claims 1 to 5, wherein the step of screening out users of the net jockey drivers from all users based on the target clustering result, and marking the net jockey drivers as the net jockey drivers, comprises:

the classes in the clusters corresponding to the target cluster number are arranged in descending order according to the characteristic values, the classes in the clusters are divided into two groups according to the arrangement order, the first group is assumed to be a net car driver, and the second group is assumed to be a passenger;

calculating the user quantity ratio in the first group and the second group, screening out two groups with the smallest absolute value of the difference value between the user quantity ratio and the actual ratio, marking the users in the first group of the two screened groups as network taxi drivers, and marking the users in the second group as passengers; the actual ratio is the driver-passenger ratio of the network about car industry which is obtained through pre-investigation.

7. A net restraint vehicle driver identification device, comprising:

8. The apparatus of claim 7, wherein the first statistics module is specifically configured to, wherein the platform usage characteristics comprise two: the sum of the longest continuous access network taxi-taking platform duration and the continuous access network taxi-taking platform duration; screening out network addresses belonging to a network taxi platform in the mobile network detail list of the current user, arranging the starting time of the network addresses belonging to the network taxi platform according to ascending order, wherein the number of entries of the network addresses belonging to the network taxi platform is n; initializing a list L= [0] for recording, and initializing an initial value of a memory variable min_last, wherein the value of the min_last is the minute at which the start time of the 1 st record is located; calculating the difference value between the minute at which the start time of the a record is located and the min_last, and if the difference value is 0, neglecting the record; let L [ -1] =l [ -1] +1, where L [ -1] represents the value of the last bit in L, if the difference is 1; if the difference value is greater than 1, inserting 0 into the last bit in L, and reassigning the min_last to be the minute of the start time of the a-th record; judging whether a is less than or equal to n; if yes, let a=a+1, and repeat the step of calculating the difference between the minute at which the start time of the a-th record is located and the min_last; if not, taking the maximum value max (L) in L and the sum (L) of each numerical value in L as the platform use characteristics of the current user on the same day; where max (L) represents the length of time for which the longest continuous access network about the vehicle platform, and sum (L) represents the sum of the lengths of time for which the continuous access network about the vehicle platform.

9. A server, comprising: a processor and a memory;

the memory stores computer-executable instructions;

the processor executing computer-executable instructions stored in the memory causes the processor to perform the net-jockey-driver identification method of any one of claims 1 to 6.

10. A computer storage medium having stored therein computer executable instructions which, when executed by a processor, implement the net restraint vehicle driver identification method of any one of claims 1 to 6.