CN107896153B - Traffic package recommendation method and device based on mobile user internet surfing behavior - Google Patents
Traffic package recommendation method and device based on mobile user internet surfing behavior Download PDFInfo
- Publication number
- CN107896153B CN107896153B CN201711128666.5A CN201711128666A CN107896153B CN 107896153 B CN107896153 B CN 107896153B CN 201711128666 A CN201711128666 A CN 201711128666A CN 107896153 B CN107896153 B CN 107896153B
- Authority
- CN
- China
- Prior art keywords
- user
- mobile
- users
- mobile user
- application
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 19
- 238000004364 calculation method Methods 0.000 claims description 8
- 239000000126 substance Substances 0.000 claims description 6
- 230000006399 behavior Effects 0.000 abstract description 17
- 230000002596 correlated effect Effects 0.000 abstract description 3
- 238000007405 data analysis Methods 0.000 abstract description 3
- 230000000875 corresponding effect Effects 0.000 description 9
- 238000004458 analytical method Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L12/00—Data switching networks
- H04L12/02—Details
- H04L12/14—Charging, metering or billing arrangements for data wireline or wireless communications
- H04L12/1485—Tariff-related aspects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0241—Advertisements
- G06Q30/0251—Targeted advertisements
- G06Q30/0255—Targeted advertisements based on user history
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/24—Accounting or billing
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Signal Processing (AREA)
- Finance (AREA)
- Computer Networks & Wireless Communication (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a flow package recommending method and a device based on the internet surfing behavior of a mobile user, which classifies the mobile user by analyzing the internet surfing behavior of the user with large data volume and utilizing the interest score of the mobile user to each application, thereby dividing users with similar access application interests into similar users, taking the flow package ordered by the users with the closest access application interests and the similar purchasing abilities to the users of packages to be recommended in the similar users as the reference basis for recommending the flow package, the method is characterized in that users with similar access interests and purchasing abilities are correlated, so that accurate pushing of flow packages is achieved, the flow packages recommended to the users are obtained based on big data analysis of user internet surfing behaviors, actual requirements of the users are met, success rate of upgrading the flow packages by the mobile users can be effectively improved, and practical economic benefits are brought to operators.
Description
Technical Field
The invention relates to the technical field of communication, in particular to a flow package recommendation method and device based on internet surfing behaviors of mobile users.
Background
The mobile internet has entered a rapid development period, and it has become a preference of users to access the network using mobile terminals. The demand of mobile users on flow is increasing, the situation that the flow in a package is insufficient often occurs, and upgrading the flow package is a main means for solving the problems. For operators, flow management is an important marketing means, so that the user off-network rate can be effectively reduced, and the operation income is increased. However, the flow packages of the operators are various, and it is difficult for the ordinary users to judge which is the most suitable one, and the ordinary users only can turn to the marketer.
The existing flow package recommendation scheme mainly depends on that marketers know user information (such as gender, identity, age and the like) by adopting a simple questionnaire mode, and recommend corresponding flow packages to mobile users in a telephone promotion mode, for example, package with high cost performance is recommended to student users, and package with high flow is recommended to business people. The conventional flow package recommendation scheme is mainly judged by subjective experience of marketers, is lack of detailed system analysis and data support, particularly support of specific services, cannot sense the real requirements of users, is difficult to truly meet the requirements of the users for flow packages recommended to the users, and is low in success rate of changing or upgrading the flow packages by mobile users.
Therefore, a traffic package recommendation scheme based on the internet access behavior of the mobile user is needed to solve the above problems.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a flow package recommendation method and device based on the internet access behavior of a mobile user, and aims to solve the problems that flow package recommendation is inaccurate and has a large difference with user requirements.
In order to solve the technical problems, the invention adopts the following technical scheme:
the invention provides a flow package recommendation method based on a mobile user internet surfing behavior, which comprises the following steps:
calculating interest scores of the mobile users to each application according to the internet surfing data of the mobile users;
classifying the mobile users according to the interest scores of the mobile users to each application;
and in the same type of mobile users, calculating the application access interest similarity between every two users, and recommending flow packages to the mobile users according to the package flow ordered by the mobile users and the application access interest similarity.
Preferably, the calculating the interest score of the mobile user for each application according to the internet surfing data of the mobile user specifically includes:
calculating the traffic utilization ratio q of the mobile user to each applicationjiAnd access duration ratio pji(ii) a Wherein the content of the first and second substances,i is an application, j is a mobile user, ujiTraffic, U, using application i for mobile user jjTraffic using all applications for mobile user j;tjiduration of use of application i for mobile user j, TjThe duration of using all applications for mobile user j;
calculating weights w for applicationsiWherein, in the step (A),uiusing the total flow of the application i for each mobile user, and using the total flow of all the applications for each mobile user;
according to qji、pji、wiAnd a preset first weight α and a preset second weight β, and calculating the interest score Q of the mobile user to the application according to the following formulaji:Qji=(α*pji+β*qji)*wiWherein α + β is 1.
Preferably, the classifying the mobile users according to the interest scores of the mobile users in each application specifically includes:
and classifying the mobile users into a plurality of categories according to the interest scores of the mobile users to the applications by adopting a clustering algorithm.
Preferably, the recommending the flow package to the mobile user according to the package flow ordered by the mobile user and the application access interest similarity specifically includes:
determining a user which is most matched with the mobile user according to package flow ordered by the mobile user and the application access interest similarity, and recommending the flow package ordered by the user which is most matched with the mobile user to the mobile user.
Preferably, the determining, according to the package flow ordered by the mobile user and the application access interest similarity, a user that is most matched with the mobile user includes:
calculating package flow difference between the mobile user and other users in the same class;
calculating a user similarity difference score between the mobile user and other users in the same class through weighted summation according to a preset third weight gamma and a preset fourth weight delta as well as package flow difference and application access interest similarity between the mobile user and other users in the same class, and determining a lowest user similarity difference score between the mobile user and other users in the same class; wherein γ + δ is 1;
and if the package flow ordered by the user corresponding to the lowest user similarity difference score is larger than the package flow ordered by the mobile user, the user corresponding to the lowest user similarity difference score is the user which is most matched with the mobile user.
The present invention also provides a server, comprising: an interest score calculation module, a user classification module and a flow package recommendation module,
the interest score calculating module is used for calculating the interest scores of the mobile users to each application according to the internet surfing data of the mobile users;
the user classification module is used for classifying the mobile users according to the interest scores of the mobile users to the applications, which are calculated by the interest score calculation module;
the flow package recommending module is used for calculating the application access interest similarity between every two users in the same type of mobile users, and recommending flow packages to the mobile users according to the package flow ordered by the mobile users and the application access interest similarity.
Preferably, the interest score calculating module is specifically configured to calculate traffic usage ratio q of the mobile user to each applicationjiAnd access duration ratio pji(ii) a Wherein the content of the first and second substances,i is an application, j is a mobile user, ujiTraffic, U, using application i for mobile user jjTraffic using all applications for mobile user j;tjiduration of use of application i for mobile user j, TjThe duration of using all applications for mobile user j; and, calculating the weight w of each applicationiWherein, in the step (A),uiusing the total flow of the application i for each mobile user, and using the total flow of all the applications for each mobile user; and according to qji、pji、wiAnd a preset first weight α and a preset second weight β, and calculating the interest score Q of the mobile user to the application according to the following formulaji:Qji=(α*pji+β*qji)*wiWherein α + β is 1.
Preferably, the user classification module is specifically configured to classify the mobile users into a plurality of categories according to the interest scores of the mobile users for the applications by using a clustering algorithm.
Preferably, the traffic package recommendation module is specifically configured to determine, according to package traffic ordered by a mobile user and application access interest similarity, a user that is most matched with the mobile user, and recommend the traffic package ordered by the user that is most matched with the mobile user to the mobile user.
Preferably, the flow package recommendation module is specifically configured to calculate a package flow difference between the mobile user and other users in the same class; calculating a user similarity difference score between the mobile user and other users in the same class through weighted summation according to a preset third weight gamma and a preset fourth weight delta as well as package flow difference and application access interest similarity between the mobile user and other users in the same class, and determining the lowest user similarity difference score between the mobile user and other users in the same class, wherein gamma + delta is 1; and when the package flow ordered by the user corresponding to the lowest user similarity difference score is larger than the package flow ordered by the mobile user, the user corresponding to the lowest user similarity difference score is the user which is most matched with the mobile user.
According to the method, the network access behavior of the user with large data volume is analyzed, the mobile user is classified by using the interest scores of the mobile user to each application, so that the users with similar access application interests are classified into the same type of users, the flow packages ordered by the users with similar purchasing ability and closest access application interests to the package users to be recommended are taken as the reference basis for recommending the flow packages, namely the users with similar application access interests and similar purchasing ability are correlated, so that the flow packages are accurately pushed, the flow packages recommended to the users are obtained based on the large data analysis of the network access behavior of the user, the actual requirements of the users are met, the success rate of upgrading the flow packages of the mobile user can be effectively improved, and practical economic benefits are brought to operators.
Drawings
Fig. 1 is a flow chart of flow package recommendation provided by an embodiment of the present invention;
FIG. 2 is a schematic diagram of user classification according to an embodiment of the present invention;
FIG. 3 is a flow chart of the determination of the best matching user according to the embodiment of the present invention;
fig. 4 is a schematic structural diagram of a server according to an embodiment of the present invention.
Detailed Description
The technical solution of the present invention will be described clearly and completely with reference to the accompanying drawings, and it should be understood that the described embodiments are a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, are within the scope of the present invention.
According to the method and the device, users with similar interest in accessing the application are divided into users of the same type by analyzing the internet surfing behavior of the users with large data volume, the relevance among the users is emphasized, the flow packages ordered by other users of the same type are recommended to the mobile users, and the flow packages are accurately pushed.
An operator deploys a DPI (Deep packet inspection) device between an SGSN (Serving GPRS Support Node) and a GGSN (Gateway GPRS Support Node) to collect internet access records of a mobile user and store the internet access records in an internet access record database. Each internet access record contains 26 fields, which mainly comprise a mobile phone number, a user type, internet access duration, a Uniform Resource Locator (URL), an application type and the like. A BSS (Business Support System) is a System in which a telecommunications carrier stores detailed information of a user, and the user information includes information such as a traffic package ordered by the user in addition to basic information of the user.
The traffic package recommendation method based on the internet surfing behavior of the mobile user is based on comprehensive analysis of data such as internet surfing records of the mobile user and traffic packages in BSS, and traffic package recommendation is carried out on the mobile user. In the embodiment of the invention, the mobile users in one month are acquired as the data statistics period, and the internet access records of each mobile user in one month are analyzed.
Table 1 is an example of a part of the user internet log extracted from the internet log data.
TABLE 1
The flow of recommending a package based on the internet access behavior of the mobile user according to the present invention is described in detail below with reference to fig. 1.
As shown in fig. 1, the flow of recommending a package based on internet access behavior of a mobile user includes the following steps:
Specifically, the traffic utilization ratio q of the mobile user to each application is calculated according to the following formula (1)ji:
Where i is an application, j is a mobile user, ujiTraffic, U, using application i for mobile user jjTraffic using all applications for mobile user j, qjiThe traffic usage duty for mobile user j to application i.
Calculating the access time length of the mobile user to each application according to the following formula (2)Ratio pji:
Wherein, tjiDuration of use of application i for mobile user j, TjDuration of use of all applications, p, for mobile user jjiThe access duration for mobile user j to application i is proportional.
Calculating weights w for applicationsiWherein, in the step (A),uithe total traffic of application i is used for each mobile user and U is the total traffic of all applications used for each mobile user.
According to qji、pji、wiAnd a preset first weight α and a preset second weight β, and calculating the interest score of the mobile user to the application according to the following formula (3):
Qji=(α*pji+β*qji)*wi; (3)
wherein α + β is 1, QjiThe interest score for mobile user j in application i.
For the sake of clear description of the technical solution of the present invention, a mobile user (user 1) with a mobile phone number "xxxx 1" in table 1 is taken as an example for explanation. The user 1 uses two applications of the WeChat and the Baidu map, the total flow of the monthly visit is 80+6+ 24-110, the total visit time is 2+0.5+ 3.5-6, the flow of the WeChat visit is 80+ 6-86, and the time is 2+ 0.5-2.5; the traffic to access the Baidu map is 24 and the duration is 3.5. Thus, user 1 accounts for the traffic usage of WeChat by q1186/110%, the duration of access to the WeChat by user 1 is in proportion to p212.5/6-42%; traffic usage ratio q of user 1 to a Baidu map1224/110-22%, the access duration of user 1 to the hundred degree map is p22And 3.5/6-58%. Weight w of WeChat1203/335-60.6%, weight w of Baidu map224/335-7.16%, Tanbao weight w 36/335-1.8%, and so on.
And 102, classifying the mobile users according to the interest scores of the mobile users to the applications.
Specifically, a clustering (K-means) algorithm may be adopted to classify the mobile users into K classes according to the interest scores of the mobile users for each application, where 1< K < N, and N is the total number of the mobile users and can be taken from experience.
K-means is a common clustering analysis algorithm, which takes K as an input parameter and divides N objects (N is the total number of mobile users) into K clusters, so that the objects in the clusters have higher similarity and the similarity between the clusters is lower. Suppose there are 5 mobile users, user a, user B, user C, user D and user E, the accessed application is WeChat, Taobao. After calculation, the interest scores of the mobile users for WeChat and Taobao are respectively expressed in a coordinate form as follows: user a (1, 9), user B (1, 7), user C (2, 8), user D (6, 1), user E (7, 2). When the K-means cluster is used, these interest scores are calculated as the characteristic values of the user. And taking K as 2, namely clustering the users into two categories, wherein the characteristic space is a two-dimensional space due to two applications, taking the value of the interest score of the mobile user on the WeChat as an X axis, and taking the value of the interest score of the mobile user on the Taobao as a Y axis. Two users are randomly selected as initial users, and a clustering diagram shown in fig. 2 can be obtained by utilizing a K-means algorithm for calculation, as shown in fig. 2, users A, B, C are classified into one category, and users D, E are classified into one category, that is, the interests of users A, B and C for application access are closer, and the interests of users D and E for application access are closer.
And 103, calculating the application access interest similarity between every two users in the same type of mobile users.
Specifically, on the user cluster coordinates obtained in step 102, the distance between every two users of the same type of users is calculated according to the euclidean distance algorithm, and the application access interest similarity S between the mobile user j and the mobile user k is represented by the distancejkThe smaller the distance between the users of the same type is, the higher the similarity of the application access interests between the users is. Taking the example shown in FIG. 2, at the location of user A, B, CIn the class, the application access interest similarity S between the user A and the user B is respectively calculatedABApplication access interest similarity S between user A and user CACApplication access interest similarity S between user B and user CBC。
The specific implementation process of calculating the distance between two users of the same type of users according to the euclidean distance algorithm belongs to the prior art, and is not described herein again.
And 104, recommending flow packages to the mobile user according to the package flow ordered by the mobile user and the similarity of the application access interests.
Specifically, according to package flow ordered by a mobile user and application access interest similarity, a user most matched with the mobile user is determined, and the flow package ordered by the user most matched with the mobile user is recommended to the mobile user. The most matched user is the user with the closest application access interest (i.e. the most similar user internet behavior) and the closest package flow ordered (i.e. the most similar purchasing power) to the user.
The process of determining the best matching user is described in more detail later in conjunction with fig. 3.
It can be seen from step 101 and step 104 that, according to the method, the internet surfing behavior of the user with large data volume is analyzed, the mobile user is classified by using the interest scores of the mobile user for each application, so that the users with similar interest in accessing the application are classified into the same type of users, the flow package ordered by the user with the similar purchasing ability and the access application interest closest to the package user to be recommended is used as the reference basis for recommending the flow package among the same type of users, that is, the users with similar access interest and purchasing ability are correlated, so that the flow package recommended to the user is obtained based on the large data analysis of the internet surfing behavior of the user, the actual demand of the user is met, the success rate of upgrading the flow package of the mobile user can be effectively improved, and economic benefits are brought to operators practically.
The following describes the determination process of the best matching user in detail with reference to fig. 3. As shown in fig. 3, determining a user that is most matched with a mobile user according to package flow ordered by the mobile user and application access interest similarity specifically includes the following steps:
Specifically, package flow ordered by each mobile user is obtained from the BSS system, and package flow difference C of each user in the same type of users is calculated respectivelyjk,Cjk=|Uj-Uk|;UjPackage flow, U, ordered for Mobile user jkPackage flow subscribed for mobile user k, CjkIs the package traffic difference between mobile user j and mobile user k.
Specifically, the user similarity difference score between the mobile user and other users in the same class is calculated through weighted summation according to a preset third weight gamma and a preset fourth weight delta, and the package flow difference and the application access interest similarity between the mobile user and other users in the same class.
Preferably, the user similarity difference score between the mobile user and other users in the same category is calculated according to the following formula (4):
Mjk=γ*Sjk+δ*Cjk; (4)
wherein γ + δ is 1, MjkAnd scoring the user similarity difference between the mobile user j and the mobile user k.
And after calculating the user similarity difference score between the mobile user and other users in the same type of users, finding out the lowest user similarity difference score.
Specifically, if the package flow ordered by the user corresponding to the lowest user similarity difference score is greater than the package flow ordered by the mobile user, the user corresponding to the lowest user similarity difference score is the user that is the most matched with the mobile user.
It should be noted that, determining the best matching user should consider not only the lowest user similarity difference score but also the package flow ordered by the user, and only the mobile user satisfying the following two conditions is considered as the best matching user:
(1) the user similarity difference score with the mobile user is lowest;
(2) and the ordered package flow is larger than that ordered by the mobile user.
Taking the example shown in FIG. 2, package flow difference C between user A and user B is calculated separately in the class of user A, B, CABAnd package flow difference C between user A and user CAB。
According to the formula (4), calculating the user similarity difference scores M between the user A and the user B respectivelyABAnd a user similarity difference score M between user A and user CACAnd determining the lowest user similarity difference score MAC(MAC<MAB)。
If the package flow ordered by the user C is larger than the package flow ordered by the user A, the user C is the user which is most matched with the user A, and the flow package ordered by the user C can be recommended to the user A. If the package flow ordered by the user C is smaller than or equal to the package flow ordered by the user A, the package flow currently ordered by the user A is already the most appropriate at the moment, and no user which is the most matched with the user A exists, other flow packages do not need to be recommended to the user A.
It can be seen from the above step 301 plus 303 that, the method determines that the best matching user can recommend the most suitable package with more flow to the mobile user, not only considering the lowest user similarity difference score, but also considering the package flow ordered by the user, so as to meet the actual demand of the user, and bring higher profit to the operator.
Based on the same technical concept, an embodiment of the present invention further provides a server, as shown in fig. 4, where the server may include: an interest score calculation module 41, a user classification module 42, and a flow package recommendation module 43.
The interest score calculating module 41 is configured to calculate interest scores of the mobile users for the applications according to the internet surfing data of the mobile users.
The user classification module 42 is configured to classify the mobile user according to the interest score of the mobile user for each application calculated by the interest score calculation module 41.
The flow package recommending module 43 is configured to calculate an application access interest similarity between every two users in the same class of mobile users, and recommend a flow package to the mobile users according to the package flow ordered by the mobile users and the application access interest similarity.
Specifically, the interest score calculating module 41 is specifically configured to calculate the traffic usage ratio q of the mobile user to each applicationjiAnd access duration ratio pji(ii) a Wherein the content of the first and second substances,i is an application, j is a mobile user, ujiTraffic, U, using application i for mobile user jjTraffic using all applications for mobile user j;tjiduration of use of application i for mobile user j, TjThe duration of using all applications for mobile user j; and, calculating the weight w of each applicationiWherein, in the step (A),uiusing the total flow of the application i for each mobile user, and using the total flow of all the applications for each mobile user; and according to qji、pji、wiAnd a preset first weight α and a preset second weight β, and calculating the interest score Q of the mobile user to the application according to the following formulaji:Qji=(α*pji+β*qji)*wiWherein α + β is 1.
Specifically, the user classification module 42 is specifically configured to classify the mobile users into a plurality of categories according to the interest scores of the mobile users for the applications by using a clustering algorithm.
Specifically, the flow package recommending module 43 is specifically configured to determine, according to the package flow ordered by the mobile user and the application access interest similarity, a user that is most matched with the mobile user, and recommend the flow package ordered by the user that is most matched with the mobile user to the mobile user.
Specifically, the flow package recommending module 43 is specifically configured to calculate a package flow difference between the mobile user and other users in the same class; calculating a user similarity difference score between the mobile user and other users in the same class through weighted summation according to a preset third weight gamma and a preset fourth weight delta as well as package flow difference and application access interest similarity between the mobile user and other users in the same class, and determining the lowest user similarity difference score between the mobile user and other users in the same class, wherein gamma + delta is 1; and when the package flow ordered by the user corresponding to the lowest user similarity difference score is larger than the package flow ordered by the mobile user, the user corresponding to the lowest user similarity difference score is the user which is most matched with the mobile user.
It will be understood that the above embodiments are merely exemplary embodiments taken to illustrate the principles of the present invention, which is not limited thereto. It will be apparent to those skilled in the art that various modifications and improvements can be made without departing from the spirit and substance of the invention, and these modifications and improvements are also considered to be within the scope of the invention.
Claims (8)
1. A traffic package recommendation method based on a mobile user internet behavior is characterized by comprising the following steps:
calculating interest scores of the mobile users to each application according to the internet surfing data of the mobile users;
classifying the mobile users according to the interest scores of the mobile users to each application;
calculating the application access interest similarity between every two users in the same type of mobile users, and recommending flow packages to the mobile users according to package flow ordered by the mobile users and the application access interest similarity;
the calculating the interest score of the mobile user to each application according to the internet surfing data of the mobile user specifically comprises the following steps:
calculating the traffic utilization ratio of mobile users to each applicationAnd access duration ratioWherein the content of the first and second substances,i is an application, j is a mobile user,traffic, U, using application i for mobile user jjTraffic using all applications for mobile user j; duration of use of application i for mobile user j, TjThe duration of using all applications for mobile user j;
calculating weights w for applicationsiWherein, in the step (A),uiusing the total flow of the application i for each mobile user, and using the total flow of all the applications for each mobile user;
2. The method of claim 1, wherein classifying the mobile user according to the mobile user's interest score for each application comprises:
and classifying the mobile users into a plurality of categories according to the interest scores of the mobile users to the applications by adopting a clustering algorithm.
3. The method of claim 1, wherein recommending traffic packages to the mobile user according to package traffic ordered by the mobile user and the application access interest similarity comprises:
determining a user which is most matched with the mobile user according to package flow ordered by the mobile user and the application access interest similarity, and recommending the flow package ordered by the user which is most matched with the mobile user to the mobile user.
4. The method of claim 3, wherein the determining the user that is most matched with the mobile user according to the package flow ordered by the mobile user and the application access interest similarity specifically comprises:
calculating package flow difference between the mobile user and other users in the same class;
calculating a user similarity difference score between the mobile user and other users in the same class through weighted summation according to a preset third weight gamma and a preset fourth weight delta as well as package flow difference and application access interest similarity between the mobile user and other users in the same class, and determining a lowest user similarity difference score between the mobile user and other users in the same class; wherein γ + δ is 1;
and if the package flow ordered by the user corresponding to the lowest user similarity difference score is larger than the package flow ordered by the mobile user, the user corresponding to the lowest user similarity difference score is the user which is most matched with the mobile user.
5. A server, comprising: an interest score calculation module, a user classification module and a flow package recommendation module,
the interest score calculating module is used for calculating the interest scores of the mobile users to each application according to the internet surfing data of the mobile users;
the user classification module is used for classifying the mobile users according to the interest scores of the mobile users to the applications, which are calculated by the interest score calculation module;
the flow package recommending module is used for calculating the application access interest similarity between every two users in the same type of mobile users, and recommending flow packages to the mobile users according to the package flow ordered by the mobile users and the application access interest similarity;
the interest score calculating module is specifically used for calculating the traffic utilization ratio of the mobile user to each applicationAnd access duration ratioWherein the content of the first and second substances,i is an application, j is a mobile user,traffic, U, using application i for mobile user jjTraffic using all applications for mobile user j; duration of use of application i for mobile user j, TjThe duration of using all applications for mobile user j; and, calculating the weight w of each applicationiWherein, in the step (A),uiusing the total flow of the application i for each mobile user, and using the total flow of all the applications for each mobile user; and, according towiAnd a preset first weight α and a preset second weight β, and calculating the interest score of the mobile user to the application according to the following formula Wherein α + β is 1.
6. The server according to claim 5, wherein the user classification module is specifically configured to classify the mobile users into a plurality of categories based on the interest scores of the mobile users for the applications using a clustering algorithm.
7. The server of claim 5, wherein the traffic package recommendation module is specifically configured to determine a user that best matches a mobile user according to package traffic subscribed by the mobile user and application access interest similarity, and recommend the traffic package subscribed by the user that best matches the mobile user to the mobile user.
8. The server of claim 7, wherein the traffic package recommendation module is specifically configured to calculate a package traffic difference between the mobile user and other users of the same class; calculating a user similarity difference score between the mobile user and other users in the same class through weighted summation according to a preset third weight gamma and a preset fourth weight delta as well as package flow difference and application access interest similarity between the mobile user and other users in the same class, and determining the lowest user similarity difference score between the mobile user and other users in the same class, wherein gamma + delta is 1; and when the package flow ordered by the user corresponding to the lowest user similarity difference score is larger than the package flow ordered by the mobile user, the user corresponding to the lowest user similarity difference score is the user which is most matched with the mobile user.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711128666.5A CN107896153B (en) | 2017-11-15 | 2017-11-15 | Traffic package recommendation method and device based on mobile user internet surfing behavior |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711128666.5A CN107896153B (en) | 2017-11-15 | 2017-11-15 | Traffic package recommendation method and device based on mobile user internet surfing behavior |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107896153A CN107896153A (en) | 2018-04-10 |
CN107896153B true CN107896153B (en) | 2020-02-21 |
Family
ID=61805330
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711128666.5A Active CN107896153B (en) | 2017-11-15 | 2017-11-15 | Traffic package recommendation method and device based on mobile user internet surfing behavior |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107896153B (en) |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110335122A (en) * | 2019-07-11 | 2019-10-15 | 中国联合网络通信集团有限公司 | Intelligent package recommendation method and system |
CN113079479B (en) * | 2020-01-06 | 2022-10-04 | 中国移动通信集团安徽有限公司 | Package recommendation method and device and computing equipment |
CN111246389B (en) * | 2020-01-09 | 2021-03-16 | 爱讯智联科技(北京)有限公司 | Intelligent network selection method, device and system based on user behaviors |
CN113965900B (en) * | 2020-07-20 | 2023-09-05 | 中国移动通信集团安徽有限公司 | Method, device, computing equipment and storage medium for dynamically expanding flow resources |
CN113254780A (en) * | 2021-06-08 | 2021-08-13 | 武汉卓尔数字传媒科技有限公司 | Information processing method and device, electronic equipment and computer storage medium |
CN114186129B (en) * | 2021-12-10 | 2023-05-26 | 中国电信股份有限公司 | Package recommendation method, package recommendation device, electronic equipment and computer readable medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103607691A (en) * | 2013-11-26 | 2014-02-26 | 中国联合网络通信集团有限公司 | Flow package recommendation method and device |
CN104518887A (en) * | 2013-09-27 | 2015-04-15 | 中国联合网络通信集团有限公司 | Package recommending method and device |
CN105389713A (en) * | 2015-10-15 | 2016-03-09 | 南京大学 | Mobile data traffic package recommendation algorithm based on user historical data |
CN106776859A (en) * | 2016-11-28 | 2017-05-31 | 南京华苏科技有限公司 | Mobile solution App commending systems based on user preference |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20010039014A1 (en) * | 2000-01-11 | 2001-11-08 | Maxygen, Inc. | Integrated systems and methods for diversity generation and screening |
CN103906027A (en) * | 2014-03-28 | 2014-07-02 | 中国联合网络通信集团有限公司 | User value evaluation method and system based on mobile user internet surfing behaviors |
-
2017
- 2017-11-15 CN CN201711128666.5A patent/CN107896153B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104518887A (en) * | 2013-09-27 | 2015-04-15 | 中国联合网络通信集团有限公司 | Package recommending method and device |
CN103607691A (en) * | 2013-11-26 | 2014-02-26 | 中国联合网络通信集团有限公司 | Flow package recommendation method and device |
CN105389713A (en) * | 2015-10-15 | 2016-03-09 | 南京大学 | Mobile data traffic package recommendation algorithm based on user historical data |
CN106776859A (en) * | 2016-11-28 | 2017-05-31 | 南京华苏科技有限公司 | Mobile solution App commending systems based on user preference |
Also Published As
Publication number | Publication date |
---|---|
CN107896153A (en) | 2018-04-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107896153B (en) | Traffic package recommendation method and device based on mobile user internet surfing behavior | |
US11847612B2 (en) | Social media profiling for one or more authors using one or more social media platforms | |
CN108763314B (en) | Interest recommendation method, device, server and storage medium | |
CN107862022B (en) | Culture resource recommendation system | |
US11122009B2 (en) | Systems and methods for identifying geographic locations of social media content collected over social networks | |
CN110795624A (en) | Analyzing a personalization framework | |
US9245252B2 (en) | Method and system for determining on-line influence in social media | |
CN106682686A (en) | User gender prediction method based on mobile phone Internet-surfing behavior | |
US20130304818A1 (en) | Systems and methods for discovery of related terms for social media content collection over social networks | |
US20130297694A1 (en) | Systems and methods for interactive presentation and analysis of social media content collection over social networks | |
US20130297581A1 (en) | Systems and methods for customized filtering and analysis of social media content collected over social networks | |
US20150032751A1 (en) | Methods and Systems for Utilizing Subject Matter Experts in an Online Community | |
CN107784066A (en) | Information recommendation method, device, server and storage medium | |
CN104866969A (en) | Personal credit data processing method and device | |
WO2007071143A1 (en) | Method and apparatus for issuing network information | |
CN109509010B (en) | Multimedia information processing method, terminal and storage medium | |
JP2008539528A (en) | Scoring user compatibility in social networks | |
CN109409928A (en) | A kind of material recommended method, device, storage medium, terminal | |
CN102591873B (en) | A kind of information recommendation method and equipment | |
EP1738524A1 (en) | Method and system for generating a population representative of a set of users of a communication network | |
CN104750760A (en) | Application software recommending method and device | |
CN105512300B (en) | information filtering method and system | |
KR101621735B1 (en) | Recommended search word providing method and system | |
CN107025567A (en) | A kind of data processing method and device | |
US20130179223A1 (en) | Method and arrangement for segmentation of telecommunication customers |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |