CN114186129A - Package recommendation method and device, electronic equipment and computer readable medium - Google Patents

Package recommendation method and device, electronic equipment and computer readable medium Download PDF

Info

Publication number
CN114186129A
CN114186129A CN202111509717.5A CN202111509717A CN114186129A CN 114186129 A CN114186129 A CN 114186129A CN 202111509717 A CN202111509717 A CN 202111509717A CN 114186129 A CN114186129 A CN 114186129A
Authority
CN
China
Prior art keywords
package
target user
historical
usage
dimension
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.)
Granted
Application number
CN202111509717.5A
Other languages
Chinese (zh)
Other versions
CN114186129B (en
Inventor
张明哲
仲籽彦
魏丫丫
洪迪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China Telecom Corp Ltd
Original Assignee
China Telecom Corp Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by China Telecom Corp Ltd filed Critical China Telecom Corp Ltd
Priority to CN202111509717.5A priority Critical patent/CN114186129B/en
Publication of CN114186129A publication Critical patent/CN114186129A/en
Application granted granted Critical
Publication of CN114186129B publication Critical patent/CN114186129B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • 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
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0631—Item recommendations
    • 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
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/40—Business processes related to the transportation industry

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Mathematical Analysis (AREA)
  • Finance (AREA)
  • General Business, Economics & Management (AREA)
  • Mathematical Physics (AREA)
  • Strategic Management (AREA)
  • Marketing (AREA)
  • Economics (AREA)
  • Mathematical Optimization (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Computational Mathematics (AREA)
  • Accounting & Taxation (AREA)
  • Pure & Applied Mathematics (AREA)
  • Development Economics (AREA)
  • Operations Research (AREA)
  • Software Systems (AREA)
  • Probability & Statistics with Applications (AREA)
  • Algebra (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Primary Health Care (AREA)
  • Tourism & Hospitality (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Telephonic Communication Services (AREA)

Abstract

The embodiment of the disclosure provides a package recommendation method, a package recommendation device, an electronic device and a readable medium, wherein the method comprises the following steps: acquiring historical package consumption of a target user; determining historical package dimension usage of the target user under at least two package dimensions for the historical package usage of the target user; determining predicted package dimension usage of the target user in each package dimension according to historical package dimension usage of the target user in the package dimension; determining the matching degree of the target user and the alternative packages according to the predicted package dimension usage of the target user under at least two package dimensions; and determining a recommended package according to the matching degree of the alternative packages so as to send the recommended package to the target user. The package recommendation method, device, electronic equipment and readable medium provided by the embodiment of the disclosure can recommend packages with high matching degree based on different dimensionalities of the user in the package.

Description

Package recommendation method and device, electronic equipment and computer readable medium
Technical Field
The present disclosure relates to the field of information technologies, and in particular, to a package recommendation method and apparatus, an electronic device, and a computer-readable medium.
Background
When a user accesses the network, a package selected subjectively is often not matched with the subsequent use habit of the user, so that the stickiness of the user is not high. The current personalized package recommendation method generally adopts a commercial recommendation algorithm to recommend a package as a commodity. The method is different from the mass data of behaviors such as clicking of the user on the E-commerce, the monthly consumption data of the user is small in scale, and the prediction accuracy is not high.
Therefore, a new package recommendation method, apparatus, electronic device and computer readable medium are needed.
The above information disclosed in this background section is only for enhancement of understanding of the background of the disclosure.
Disclosure of Invention
In view of this, the present disclosure provides a package recommendation method, an apparatus, an electronic device, and a computer-readable medium, which can recommend a package with a high matching degree based on usage of different dimensions of a user in the package.
Additional features and advantages of the disclosure will be set forth in the detailed description which follows, or in part will be obvious from the description, or may be learned by practice of the disclosure.
According to a first aspect of an embodiment of the present disclosure, a package recommendation method is provided, where the method includes: acquiring historical package consumption of a target user; determining historical package dimension usage of the target user under at least two package dimensions for the historical package usage of the target user; determining predicted package dimension usage of the target user in each package dimension according to historical package dimension usage of the target user in the package dimension; determining the matching degree of the target user and the alternative packages according to the predicted package dimension usage of the target user under at least two package dimensions; and determining a recommended package according to the matching degree of the alternative packages so as to send the recommended package to the target user.
In an exemplary embodiment of the present disclosure, the package dimensions include two or more of the following: tariff, traffic volume, voice volume, broadband cost and broadband rate.
In an exemplary embodiment of the present disclosure, determining a predicted package dimension usage by the target user in each package dimension based on historical package dimension usage by the target user in that package dimension comprises: if the package dimension comprises the tariff, determining the predicted tariff of the target user according to the historical tariff time sequence in the historical package usage; if the package dimension comprises flow, determining the predicted flow consumption of the target user according to a historical flow consumption time sequence in the historical package consumption; and if the package dimension comprises the voice consumption, determining the predicted voice consumption of the target user according to the historical voice consumption time sequence in the historical package consumption.
In an exemplary embodiment of the present disclosure, determining the predicted tariff for the target user based on a historical time series of tariffs in the historical package volume comprises: performing stationarity detection and white noise detection on the historical tariff time sequence; and if the stationarity detection result is a weak stationary sequence and the white noise detection result is a non-white noise sequence, processing the historical tariff time sequence by adopting an ARIMA time sequence prediction model to obtain the predicted tariff of the target user.
In an exemplary embodiment of the present disclosure, the method further comprises: acquiring a training sample set, wherein the training sample set comprises a sample tariff time sequence; processing the sample tariff time sequence through the ARIMA time sequence prediction model to obtain a sample prediction tariff; and adjusting parameters of the ARIMA time sequence prediction model according to the sample tariff time sequence and the sample prediction tariff to obtain the trained ARIMA time sequence prediction model.
In an exemplary embodiment of the present disclosure, further comprising: determining a switching user for switching to a recommended package; integrating the historical package usage of the switching user into an update sample set; and optimizing parameters of the ARIMA time series prediction model according to the updated sample set.
In an exemplary embodiment of the present disclosure, determining a matching degree of the target user and an alternative package according to predicted package dimension usage amounts of the target user in at least two package dimensions includes: calculating an absolute value of a difference value between predicted package dimension usage of the target user in each package dimension and candidate package dimension usage of the candidate package in the package dimension, and obtaining an initial distance between the target user and the candidate package in each package dimension; calculating the reciprocal of the sum of the initial distance between the target user and the alternative package in each package dimension and the distance compensation, and determining the matching distance between the target user and the alternative package in each package dimension; and determining the matching degree of the target user and the alternative package according to the weighted summation result of the matching distances between the target user and the alternative package in at least two package dimensions.
According to a second aspect of the embodiments of the present disclosure, there is provided a package recommendation apparatus, including: the historical data acquisition module is used for acquiring the historical package consumption of the target user; a package dimension quantifying module, configured to determine, for the historical package usage of the target user, historical package dimension usage of the target user in at least two package dimensions; a package dimension prediction module, configured to determine, according to historical package dimension usage of the target user in each package dimension, predicted package dimension usage of the target user in the package dimension; the matching degree calculation module is used for determining the matching degree of the target user and the alternative packages according to the predicted package dimension usage of the target user under at least two package dimensions; and the package recommending module is used for determining a recommended package according to the matching degree of the alternative packages so as to send the recommended package to the target user.
According to a third aspect of the embodiments of the present disclosure, an electronic device is provided, which includes: one or more processors; storage means for storing one or more programs; when executed by the one or more processors, cause the one or more processors to implement the package recommendation method of any of the above.
According to a fourth aspect of embodiments of the present disclosure, a computer-readable medium is proposed, on which a computer program is stored, which program, when being executed by a processor, is adapted to carry out the package recommendation method according to any one of the above.
According to package recommendation methods, devices, electronic devices, and computer-readable media provided by some embodiments of the present disclosure, historical package usage by a target user is based; predicting package dimension usage of a target user under each package dimension by using historical package dimension usage of the target user under at least two package dimensions; the method can be used for carrying out refinement prediction on the use habits of users on the basis of different package dimensions, determining the matching degree of a target user and alternative packages according to the package dimension usage predicted by the target user under at least two package dimensions, integrating the refinement prediction results under a plurality of package dimensions, calculating the matching degree of each alternative package and the target user on the basis of the alternative package dimension usage of each alternative package under the at least two package dimensions, and then accurately positioning the recommended packages with high matching degree with the target user according to the matching degree of the alternative packages.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure. The drawings described below are merely some embodiments of the present disclosure, and other drawings may be derived from those drawings by those of ordinary skill in the art without inventive effort.
Fig. 1 is a system block diagram illustrating a package recommendation method and apparatus according to an exemplary embodiment.
FIG. 2 is a flow diagram illustrating a package recommendation method according to an exemplary embodiment.
FIG. 3 is a flow chart illustrating a package recommendation method according to another exemplary embodiment.
FIG. 4 is a flowchart illustrating a package recommendation method according to yet another exemplary embodiment.
FIG. 5 is a flowchart illustrating a package recommendation method according to yet another exemplary embodiment.
FIG. 6 is a schematic diagram illustrating a data acquisition flow according to an exemplary embodiment.
FIG. 7 is a schematic diagram illustrating a data processing prediction flow according to an exemplary embodiment.
FIG. 8 is a graph illustrating the predicted effect of a flow prediction model according to an exemplary embodiment.
FIG. 9 is a diagram illustrating the predictive effects of a speech prediction model according to an exemplary embodiment.
FIG. 10 is a diagram illustrating the predictive effect of a tariff prediction model according to an exemplary embodiment.
Fig. 11 is a diagram illustrating the effect of pushing a short message for recommending packages according to an exemplary embodiment.
Fig. 12 is an online business hall push effect diagram illustrating a recommended package according to an example embodiment.
Fig. 13 is a schematic diagram illustrating a handover data collection procedure according to an example embodiment.
FIG. 14 is a block diagram illustrating a package recommendation device according to an exemplary embodiment.
Fig. 15 schematically illustrates a block diagram of an electronic device in an exemplary embodiment of the disclosure.
Detailed Description
Example embodiments will now be described more fully with reference to the accompanying drawings. Example embodiments may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of example embodiments to those skilled in the art. The same reference numerals denote the same or similar parts in the drawings, and thus, a repetitive description thereof will be omitted.
The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to provide a thorough understanding of embodiments of the invention. One skilled in the relevant art will recognize, however, that the invention may be practiced without one or more of the specific details, or with other methods, components, devices, steps, and so forth. In other instances, well-known methods, devices, implementations or operations have not been shown or described in detail to avoid obscuring aspects of the invention.
The drawings are merely schematic illustrations of the present invention, in which the same reference numerals denote the same or similar parts, and thus, a repetitive description thereof will be omitted. Some of the block diagrams shown in the figures do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different networks and/or processor devices and/or microcontroller devices.
The flow charts shown in the drawings are merely illustrative and do not necessarily include all of the contents and steps, nor do they necessarily have to be performed in the order described. For example, some steps may be decomposed, and some steps may be combined or partially combined, so that the actual execution sequence may be changed according to the actual situation.
The following detailed description of exemplary embodiments of the invention refers to the accompanying drawings.
Fig. 1 is a system block diagram illustrating a package recommendation method and apparatus according to an exemplary embodiment.
In the system 100 of package recommendation method and apparatus, the server 105 may be a server providing various services, such as a background management server (for example only) providing support for a package recommendation system operated by a user using the terminal devices 101, 102, 103 through the network 104. The background management server may analyze and perform other processing on the received package recommendation request and other data, and feed back a processing result (for example, a package recommendation, a matching degree of the target user and the alternative package-only an example) to the terminal device.
The server 105 may be a server of one entity, and may also be composed of a plurality of servers, for example, a part of the server 105 may be, for example, used as a package recommendation task submitting system in the present disclosure, and is used to obtain a task to be executed with a package recommendation command; and a portion of server 105 may also be used, for example, as a package recommendation system in the present disclosure, for obtaining historical package usage of a target user; determining historical package dimension usage of the target user under at least two package dimensions for the historical package usage of the target user; determining predicted package dimension usage of the target user in each package dimension according to historical package dimension usage of the target user in the package dimension; determining the matching degree of the target user and the alternative packages according to the predicted package dimension usage of the target user under at least two package dimensions; and determining a recommended package according to the matching degree of the alternative packages so as to send the recommended package to the target user.
FIG. 2 is a flow diagram illustrating a package recommendation method according to an exemplary embodiment. The package recommendation method provided by the embodiments of the present disclosure may be executed by any electronic device with computing processing capability, for example, the terminal devices 101, 102, and 103 and/or the server 105, and in the following embodiments, the server executes the method as an example for illustration, but the present disclosure is not limited thereto. The package recommendation method provided by the embodiment of the disclosure may include steps S202 to S210.
As shown in FIG. 2, in step S202, the historical package usage of the target user is obtained.
In step S204, historical package dimension usage of the target user in at least two package dimensions is determined for the historical package usage of the target user.
In embodiments of the present disclosure, a package dimension of historical package volume may have multiple dimensions, such as a tariff dimension, a traffic volume dimension, a voice volume dimension, a broadband cost dimension, and a broadband rate dimension. Package dimensions include two or more of the following: tariff, traffic volume, voice volume, broadband cost and broadband rate. When a certain dimension has a restriction condition, for example, the traffic volume dimension has a regional restriction and is divided into a local traffic volume and a national traffic volume, the traffic volume dimension may be further divided into the local traffic volume dimension and the national traffic volume dimension. As another example, the flow volume dimension may be further divided into a directional flow volume, an in-jacket flow volume, and an out-of-jacket flow volume. The directional flow has, for example, a limiting condition: only for a specific application. The in-house flow has, for example, the limiting condition: traffic usage of preferred usage contained within the package. The sheath flow rate has, for example, a limiting condition: the flow rate in the sleeve is 0 and the usable flow rate is used. For another example, when the voice usage dimension has a restriction of a dial object and is divided into a home group call voice dimension, a local voice call voice dimension, and a national voice call voice dimension, the voice usage dimension may be further divided into a home group call voice dimension, a local voice call voice dimension, and a national voice call voice dimension.
The current package information of the target user can be obtained, and the current package information can comprise current package cost, current package flow, current package language volume, current package broadband cost, current package broadband rate and the like. For example, when the current package tariff of the target user is less than the historical package tariff in the historical package usage, and/or the current package flow of the target user is less than the historical flow usage in the historical package usage, and/or the current package speech volume of the target user is less than the historical speech usage in the historical package usage, and/or the current package broadband fee of the target user is less than the historical package broadband usage in the historical package usage, and/or the current package broadband rate of the target user is less than the historical package broadband rate in the historical package usage, steps S202 to S210 of this embodiment are performed again.
In step S206, a predicted package dimension usage of the target user in each package dimension is determined according to the historical package dimension usage of the target user in the package dimension.
In the embodiment of the present disclosure, if the package dimension includes a tariff, the predicted tariff of the target user may be determined according to a historical tariff time sequence in the historical package usage;
if the package dimension comprises flow, determining the predicted flow consumption of the target user according to a historical flow consumption time sequence in the historical package consumption;
and if the package dimension comprises the voice consumption, determining the predicted voice consumption of the target user according to the historical voice consumption time sequence in the historical package consumption.
The historical tariff time series may be a series generated according to the tariff of the target user in each month in a time sequence. The historical traffic volume time series may be a series generated in chronological order based on the traffic volume of the target user in each month. The historical speech volume time series may be a series generated in chronological order based on the target user's speech volume in each month.
In step S208, the matching degree between the target user and the alternative package is determined according to the predicted package dimension usage amount of the target user in at least two package dimensions.
The alternative package may be, for example, obtained by recalling user information of the target user, and the alternative package may be a package currently available for the recalled target user. After the alternative packages are obtained, package information for each alternative package may be extracted, such as alternative package dimension usage at each package dimension.
The target distance between the predicted package dimension usage of the target user in each package dimension and the candidate package dimension usage of the candidate package in the package dimension can be determined, and the matching degree between the target user and the candidate package is determined according to the weighted calculation result of the target distance in each package dimension.
In step S210, a recommended package is determined according to the matching degree of the alternative packages, so as to send the recommended package to the target user.
And the alternative packages can be sorted in a descending order according to the matching degree, the first N alternative packages in the sorting result are used as recommended packages and sent to the target user, and N is an integer greater than 0.
According to the package recommendation method provided by the embodiment of the disclosure, based on the historical package usage of the target user; predicting package dimension usage of a target user under each package dimension by using historical package dimension usage of the target user under at least two package dimensions; the method can be used for carrying out refinement prediction on the use habits of users on the basis of different package dimensions, determining the matching degree of a target user and alternative packages according to the package dimension usage predicted by the target user under at least two package dimensions, integrating the refinement prediction results under a plurality of package dimensions, calculating the matching degree of each alternative package and the target user on the basis of the alternative package dimension usage of each alternative package under the at least two package dimensions, and then accurately positioning the recommended packages with high matching degree with the target user according to the matching degree of the alternative packages.
Further, after the recommended package is sent to the target user, a switching user for switching to the recommended package can be determined; integrating the historical package usage of the switching user into an update sample set; and optimizing parameters of the ARIMA time series prediction model according to the updated sample set.
When a target user switches a package to a recommended package, the predicted package dimension usage of the target user is accurate, parameters of the ARIMA time sequence prediction model are further optimized based on the historical package usage of the switching user, and accurate prediction of the predicted package usage dimension can be achieved.
FIG. 3 is a flow chart illustrating a package recommendation method according to another exemplary embodiment. In the package recommendation method according to the embodiment of the present disclosure, if the package dimension includes a tariff, determining the predicted tariff time of the target user according to the historical tariff time sequence in the historical package usage may include steps S302 to S304.
As shown in fig. 3, in step S302, stationarity detection and white noise detection are performed on the historical tariff time series.
In the embodiment of the disclosure, stationarity detection may be performed first, and white noise detection may be performed when the stationarity detection result is a weak stationary sequence. Stationarity detection requires that the fitted curve obtained via the sample time series continue "inertially" following the existing morphology for a future period. White noise detection may be considered when the stationarity detection result is a weak stationary sequence. Not all stationary sequences are worth modeling. Only sequences with close correlation between sequence values and certain influence of historical data on future development are worth mining effective information in the historical data to predict the future development of the sequences.
In step S304, if the stationarity detection result is a weak stationary sequence and the white noise detection result is a non-white noise sequence, the ARIMA time sequence prediction model is used to process the historical tariff time sequence to obtain the predicted tariff of the target user.
Among them, the difference integration Moving Average Autoregressive model (ARIMA) is also called integration Moving Average Autoregressive model (Moving may also be called sliding). In ARIMA (p, d, q), AR is "autoregressive" and p is the number of autoregressive terms; MA is "moving average", q is the number of terms of the moving average, and d is the number of differences (order) made to make it a stationary sequence. The term "difference", although not shown in the english name of ARIMA, is a critical step.
The ARIMA time series prediction model can be obtained by the following steps:
acquiring a training sample set, wherein the training sample set comprises a sample tariff time sequence; processing the sample tariff time sequence through the ARIMA time sequence prediction model to obtain a sample prediction tariff; and adjusting parameters of the ARIMA time sequence prediction model according to the sample tariff time sequence and the sample prediction tariff to obtain the trained ARIMA time sequence prediction model.
Further, if the package dimension includes flow, when the predicted flow usage of the target user is determined according to a historical flow usage time sequence in the historical package usage, stationarity detection and white noise detection can be performed on the historical flow usage time sequence; and if the stationarity detection result is a weak stationary sequence and the white noise detection result is a non-white noise sequence, processing the historical flow usage time sequence by adopting an ARIMA time sequence prediction model to obtain the predicted flow usage of the target user.
Further, if the package dimension includes the voice usage, stationarity detection and white noise detection can be performed on the historical voice usage time sequence when the predicted voice usage of the target user is determined according to the historical voice usage time sequence in the historical package usage; and if the stationarity detection result is a weak stationary sequence and the white noise detection result is a non-white noise sequence, processing the historical voice usage time sequence by adopting an ARIMA time sequence prediction model to obtain the predicted voice usage of the target user.
FIG. 4 is a flowchart illustrating a package recommendation method according to yet another exemplary embodiment. The package recommendation method provided by the embodiment of the disclosure may include steps S402 to S406.
As shown in fig. 4, in step S402, an absolute value of a difference between the predicted package dimension usage of the target user in each package dimension and the candidate package dimension usage of the candidate package in the package dimension is calculated, and an initial distance between the target user and the candidate package in each package dimension is obtained.
In the embodiment of the disclosure, when the package dimension includes the tariff, the initial distance between the target user and the alternative package in the tariff dimension may be determined according to the absolute value of the difference between the predicted tariff of the target user and the package tariff of the alternative package.
When the package dimension includes the flow, the initial distance between the target user and the alternative package in the flow dimension can be determined according to the absolute value of the difference value between the predicted flow usage of the target user and the package flow of the alternative package.
When the package dimension includes the voice usage, the initial distance between the target user and the alternative package in the voice usage dimension can be determined according to the absolute value of the difference between the predicted voice usage of the target user and the package voice volume of the alternative package.
In step S404, the inverse of the sum of the initial distance and the distance compensation of the target user to the alternative package in each package dimension is calculated, and the matching distance of the target user to the alternative package in each package dimension is determined.
The compensation distance is a predetermined value, and in one embodiment, the value of the compensation distance may be 1.
When the package dimension includes a tariff, the matching distance of the target user to the alternative package in the tariff dimension may be represented as:
Figure BDA0003405285850000101
when the package dimension includes traffic, the matching distance of the target user to the alternative package in the traffic dimension can be expressed as:
Figure BDA0003405285850000102
when the package dimension includes voice usage, the matching distance of the target user to the alternative package in the voice usage dimension may be represented as:
Figure BDA0003405285850000103
in step S406, the matching degree between the target user and the alternative package is determined according to the weighted summation result of the matching distances between the target user and the alternative package in at least two package dimensions.
When the package dimension includes the tariff, the flow and the voice consumption, the matching degree of the target user and the alternative package can be expressed as:
Figure BDA0003405285850000111
and s is the matching degree of the target user and the alternative package. Omega1、ω2、ω3Is a weight value.
In this embodiment, based on the matching distance between the target user and the alternative package in each package dimension, the matching degree between the target user and the alternative package can be accurately located.
Further, the weight value can be iteratively updated by calculating the switching user and the recommended package selected by the switching user.
FIG. 5 is a flowchart illustrating a package recommendation method according to yet another exemplary embodiment. The package recommendation method provided by the embodiment of the disclosure may include steps S502 to S512.
As shown in fig. 5, in step S502, data is extracted.
For example, after detecting that the target user logs in an online channel (e.g., an online business hall), data such as the user identity and historical consumption of the target user can be extracted according to the user account, and data cleaning and feature processing are completed. In addition, packages that the user can handle can be recalled and package information extracted according to the user information of the target user.
In this step, after the target user logs in the online channel, the login account of the target user is collected, and the user information is associated, including but not limited to:
1) current package information, for example: package amount, in-package flow, in-package voice, etc.;
2) historical package usage, for example: the method comprises the steps of paying off charge in a period of about M months, flow usage in a period of about M months, voice usage in a period of about M months and the like, wherein M is an integer larger than 0.
Secondly, recalling packages that the target user can handle as alternative packages, and extracting information under each package dimension, including but not limited to: package amount, in-package flow, in-package voice, etc.
Detailed steps can be seen in fig. 6, as shown in fig. 6, in the data collection process, the online channel background uses a front-end point-burying technology to write user login data into a local log (which may include one or more of a login log, a browsing log, and a consumption log) in real time. Data association: and collecting the log data into middleware (such as filebeat) in real time, and performing real-time calculation. Data storage: and writing the data after real-time calculation into various databases for storage, and regarding the users as target users to be recommended.
The tag library can be queried in real time through the user identification of the target user, and user information of the user required by the prediction model is obtained.
In step S504, the monthly usage of the target user is predicted based on the user data modeling of the target user.
The user data is historical package consumption, for example, the package can be divided into 3 package dimensions of tariff, flow consumption and voice consumption, and the monthly consumption of the target user in each package dimension is modeled and predicted respectively.
Three consumption attributes of flow, voice and charge in the historical package consumption of the target user can be used, and an ARIMA time series prediction model is used for modeling and predicting the monthly consumption of the user respectively.
The detailed steps are as follows:
historical data of the associated target user, such as charging system data, user behavior data, pushing data, handling data, clustering label data and the like, are obtained to obtain flow volume, voice volume and tariff data of the user, and three types of models are obtained through characteristic engineering and training, wherein the specific process is shown in figure 7. In FIG. 7, HDFS, ES, MySQL, REDIS are databases.
As shown in fig. 7, (1) a time series prediction model of a ARIMA (differential autoregressive mobile evaluation model) is constructed based on a package dimensions of traffic volume, voice volume, tariff, and the like. A is an integer greater than 0, and this embodiment is exemplified by a being 3, that is, package dimensions include traffic volume, voice volume, and tariff.
Firstly, cleaning the collected data to obtain regular data.
Secondly, preprocessing the data in a time sequence: stationarity detection and white noise test.
a. Stationarity detection
The testing of the stationarity of the data is an important step in time series analysis, and the stationarity is the requirement that the fitted curve obtained by the sample time series can continue along the existing form inertially in a future period.
Smoothness detection can be detected by looking at the timing diagram and ADF unit root. If there is a unit root, the sequence is not smoothed, and it is necessary to smooth the sequence, and perform logarithmic transformation, exponential smoothing, differentiation, and decomposition in this order.
And (4) conclusion: the mean value is constant, the variance exists all the time, the autocovariance does not fluctuate along with time, and the self-covariance can be regarded as a weak stable sequence to meet the prediction condition.
b. White noise detection
Not all stationary sequences are worth modeling. Only sequences with close correlation among sequence values and historical data having certain influence on future development are worth mining effective information in the historical data to predict the future development of the sequences.
After calculating the p-value of the Q statistic based on the Ljung-Box Q statistic, if the p-value is greater than 0.05, it is indicated that the time series is a white noise series.
And (4) conclusion: pvalue statistic, 0.0387<0.05, based on chi-square distribution, so the original hypothesis can be rejected, considering the sequence not to be a white noise sequence.
Model identification order determination
In ARIMA (p, d, q), AR is "autoregressive" and p is the number of autoregressive terms; i is the difference, d is the number of differences (order) made to make it a plateau sequence; MA is "moving average", and q is the number of terms of the moving average. The ACF autocorrelation coefficient can determine the value of q, and the PACF partial autocorrelation coefficient can determine the value of q. ARIMA principle: a model established by converting a non-stationary time series to a stationary time series and then regressing the dependent variable only for its lag value and the present and lag values of the random error term.
A. Determining regression terms
Autoregressive model (AR): describing the relation between the current value and the historical value, and predicting the variable by using the historical time data of the variable.
Formula for the p-order autoregressive process:
Figure BDA0003405285850000131
wherein e istIs whiteNoise.
PACF, partial autocorrelation function (determining p value), eliminates the correlation degree of x (t-k) on x (t) after interference of k-1 intermediate random variables x (t-1), x (t-2), … … and x (t-k + 1).
B. Determining a moving average number of terms
Moving average Model (MA): the moving average model focuses on the accumulation of error terms in the autoregressive model, and the moving average method can effectively eliminate random fluctuation in prediction.
The formula definition of the q-order autoregressive process:
Figure BDA0003405285850000141
in ACF, the autocorrelation function (determining the q value) reflects the correlation between values of the same sequence at different time sequences. x (t) is simultaneously influenced by intermediate k-1 random variables x (t-1), x (t-2), … … and x (t-k +1), and the k-1 random variables have correlation with x (t-k), so that the autocorrelation coefficient p (k) is actually doped with the influence of other variables on x (t) and x (t-k)
Figure BDA0003405285850000142
C. Determining number of differences
Autoregressive moving average model (ARMA): combination of autoregressive and moving average
Figure BDA0003405285850000143
Establishing a model:
a. the first type of model: ARIMA flow prediction model
Obtaining model optimal parameters through model training:
regression term: 4
Moving average term: 5
Difference times are as follows: 2
ARIMA(4,2,5)
As shown in fig. 8, the prediction effect graph is an ARIMA traffic prediction model.
b. The second type of model: ARIMA speech prediction model
Obtaining model optimal parameters through model training:
regression term: 2
Moving average term: 2
Difference times are as follows: 4
ARIMA(2,4,2)
FIG. 9 is a diagram of prediction effect for ARIMA speech prediction model
c. The third type of model: ARIMA tariff prediction model
Obtaining model optimal parameters through model training:
regression term: 3
Moving average term: 2
Difference times are as follows: 2
ARIMA(3,2,2)
As shown in FIG. 10, it is a prediction effect diagram of ARIMA tariff prediction model
(2) Model evaluation
And (3) evaluating a regression model: and extracting consumption records of B months in the history of the user, wherein B is an integer larger than 0, predicting data of one month by using an ARIMA model, then comparing the predicted data with real data to calculate an evaluation index, and judging the quality of the model mainly through an R square value.
For each model, extracting 1000 general-purpose consumer historical consumption records of one year, iteratively training the model, and adjusting for many times, wherein the optimal model parameters and evaluation results are as follows:
a. flow usage prediction model assessment
According to the optimal parameters of the model: regression term (4), moving average term number (5) and difference degree (2). Evaluation of the resulting model: mean absolute error (6.2), mean variance (9.6), R-squared value (0.84). The evaluation results were: the model fit is good.
b. Speech usage prediction model evaluation
According to the optimal parameters of the model: regression term (2), moving average term number (2), and difference degree (4). Evaluation of the resulting model: mean absolute error (4.5), mean variance (11.2), R-squared value (0.88). The evaluation results were: intact model fitting
c. Tariff usage prediction model assessment
According to the optimal parameters of the model: regression term (3), moving average term number (2), and difference degree (2). Evaluation of the resulting model: mean absolute error (5.8), mean variance (10.4), R-squared value (0.85). The evaluation results were: intact model fitting
In step S506, the candidate package matching degree is evaluated according to the monthly usage combination of the user.
According to the monthly predicted flow consumption, the predicted voice consumption and the predicted tariff of the target user, defining formula weighted combination, calculating the matching degree of the target user and the alternative package, obtaining the most matched alternative package Top N based on the matching degree sorting, and determining the best matched alternative package as the recommended package. The calculation manner of the matching degree of the alternative package can be seen in the related content of step S406.
In step S508, the detailed information of the recommended package, the preferential event, and the like are recommended to the target user.
And recommending the N pieces of predicted recommended package information to the target user in real time.
Recommendation channels are for example: recommending the short messages, the online business hall client and the like to the user in real time, wherein the details comprise the comparison details of the existing package and the optimal package;
take short message and online business hall client push as an example:
1) short message channel
Firstly, a short message interface is sent by adopting a short message platform: and/tymh _ interface _ sms/send.
Step two, ginseng introduction: destMobiles (mobile phone number array), content (short message content), sysCode (initiator platform code), mac (signature field), sendDate (sending time), smsType (short message type).
Responding: code (response code), errorescription (response description), phone (telephone number), flag (short message sending identifier), identifier (short message unique identification code), smsId (short message code).
Transmitting a restriction rule: the same user transmits once per week.
Fifth, the content of the short message template is as follows: honored xxx users, your current package is: xxx. In view of your package use, the following packages are more suitable for you: xxx packages containing xx traffic, xx voices. If you switch to the package, you will save x dollars per month. Poke me to see the detail xxx address.
Fig. 11 shows the effect of pushing the sms message.
2) Online business hall client channel
Adopting a data platform api pushing interface: /ipush/api/sync/push.
Step two, ginseng introduction: destination (phone number), taskId (API task id), third Access token, Extendinfo (extended information), MessageType, pushWay (push method).
Responding: code (response code), errorMessage (response description), pushId (push id), messageId (message id), msgid (push state id).
Transmitting a restriction rule: the same user transmits once a day.
Template information: small principal! Matching to a better package xxx, poking me to see the details.
As shown in fig. 12, it is a diagram of the push effect of the online business hall.
In step S510, package switching information of the target user is acquired.
Obtaining the use detail information such as the current flow usage amount and the current voice usage amount of the target user, and the information such as whether the target user switches packages, including but not limited to switching time, package price, etc.
And collecting data whether the user performs switching or not, and storing the data in an off-line data bin for system optimization. The user data that allows switching is written in the database, and the user's package is switched in the next month, as shown in fig. 13.
In step S512, effect feedback and system optimization are performed.
And acquiring information such as actual flow, voice and app preference of the target user, using the information as historical data to adjust the model parameters in the step S504, and optimizing the system effect.
The package recommendation scheme provided by the embodiment has the following technical effects. Firstly, feature selection and innovation on a model: different from the traditional method of predicting packages as commodities, a plurality of ARIMA time sequence prediction models are constructed according to the inherent attribute characteristics of the packages, so that the user habits are independently predicted in various types, and the accuracy is remarkably improved. Secondly, package recommendation accuracy is improved: the recommendation method provided by the text divides the inherent attributes of the package and predicts the monthly usage of the user respectively, and due to the fact that the monthly traffic and the voice usage of the user are extremely habituated, the user habits are predicted independently, and accuracy is improved greatly. And then selecting the optimal package through the weighted combination of the monthly usage of the user obtained through prediction.
It should be clearly understood that this disclosure describes how to make and use particular examples, but the principles of this disclosure are not limited to any details of these examples. Rather, these principles can be applied to many other embodiments based on the teachings of the present disclosure.
Those skilled in the art will appreciate that all or part of the steps for implementing the above embodiments are implemented as a computer program executed by a Central Processing Unit (CPU). When executed by a central processing unit CPU, performs the above-described functions defined by the above-described methods provided by the present disclosure. The program may be stored in a computer readable storage medium, which may be a read-only memory, a magnetic or optical disk, or the like.
Furthermore, it should be noted that the above-mentioned figures are only schematic illustrations of the processes involved in the methods according to exemplary embodiments of the present disclosure, and are not intended to be limiting. It will be readily understood that the processes shown in the above figures are not intended to indicate or limit the chronological order of the processes. In addition, it is also readily understood that these processes may be performed synchronously or asynchronously, e.g., in multiple modules.
The following are embodiments of the disclosed apparatus that may be used to perform embodiments of the disclosed methods. For details not disclosed in the embodiments of the apparatus of the present disclosure, refer to the embodiments of the method of the present disclosure.
FIG. 14 is a block diagram illustrating a package recommendation device according to an exemplary embodiment. Referring to fig. 14, a package recommendation apparatus 1400 provided in an embodiment of the present disclosure may include: historical data acquisition module 1402, package dimension quantification module 1404, package dimension prediction module 1406, matching degree calculation module 1408, and package recommendation module 1410.
In package recommendation apparatus 1400, a historical data acquisition module 1402 may be used to acquire historical package usage of a target user.
Package dimension quantification module 1404 may be operative to determine historical package dimension usage by the target user in at least two package dimensions for the historical package usage by the target user.
Package dimension prediction module 1406 may be used to determine a predicted package dimension usage by the target user in each package dimension based on historical package dimension usage by the target user in that package dimension.
Matching degree calculation module 1408 may be configured to determine a degree of matching of the target user with an alternative package based on predicted package dimension usage by the target user in at least two package dimensions.
Package recommending module 1410 may be configured to determine a recommended package according to the matching degree of the alternative packages, so as to send the recommended package to the target user.
According to the package recommending device provided by the embodiment of the disclosure, based on the historical package usage of the target user; predicting package dimension usage of a target user under each package dimension by using historical package dimension usage of the target user under at least two package dimensions; the method can be used for carrying out refinement prediction on the use habits of users on the basis of different package dimensions, determining the matching degree of a target user and alternative packages according to the package dimension usage predicted by the target user under at least two package dimensions, integrating the refinement prediction results under a plurality of package dimensions, calculating the matching degree of each alternative package and the target user on the basis of the alternative package dimension usage of each alternative package under the at least two package dimensions, and then accurately positioning the recommended packages with high matching degree with the target user according to the matching degree of the alternative packages.
In an exemplary embodiment, the package dimensions may include two or more of the following: tariff, traffic and voice usage.
In an exemplary embodiment, package dimension prediction module 1406 may include: the price dimension prediction unit can be used for determining the predicted price of the target user according to the historical price time sequence in the historical package amount if the package dimension comprises the price; a flow dimension prediction unit, configured to determine, if the package dimension includes a flow, a predicted flow usage of the target user according to a historical flow usage time sequence in historical package usage; and the voice dimension prediction unit can be used for determining the predicted voice consumption of the target user according to the historical voice consumption time sequence in the historical package consumption if the package dimension comprises the voice consumption.
In an exemplary embodiment, the tariff dimension prediction unit may include: the time sequence detection subunit is used for performing stationarity detection and white noise detection on the historical tariff time sequence; and the charge dimension prediction subunit is used for processing the historical charge time sequence by adopting an ARIMA time sequence prediction model to obtain the predicted charge of the target user if the stationarity detection result is a weak stationary sequence and the white noise detection result is a non-white noise sequence.
In an exemplary embodiment, the package recommendation apparatus provided in the embodiments of the present disclosure may further include: the system comprises a sample set acquisition module, a sample set acquisition module and a training sample set acquisition module, wherein the sample set acquisition module can be used for acquiring a training sample set, and the training sample set comprises a sample tariff time sequence; the sample set processing module can be used for processing the sample expense time sequence through the ARIMA time sequence prediction model to obtain a sample prediction expense; and the model training unit can be used for adjusting parameters of the ARIMA time sequence prediction model according to the sample tariff time sequence and the sample prediction tariff to obtain the trained ARIMA time sequence prediction model.
In an exemplary embodiment, the package recommendation apparatus provided in the embodiments of the present disclosure may further include: the switching user positioning module can be used for determining a switching user for switching to the recommended package; an update sample set module, configured to integrate the historical package usage of the handover user into an update sample set; an iterative update module operable to optimize parameters of the ARIMA time series prediction model according to the updated sample set.
In an exemplary embodiment, the degree of match calculation module 1408 may include: an initial distance calculating unit, configured to calculate an absolute value of a difference between predicted package dimension usage by the target user in each package dimension and candidate package dimension usage by the candidate package in the package dimension, and obtain an initial distance between the target user and the candidate package in each package dimension; a matching distance calculation unit, configured to calculate an inverse of a sum of an initial distance between the target user and the alternative package in each package dimension and a distance compensation value, and determine a matching distance between the target user and the alternative package in each package dimension; and the matching degree calculation unit can be used for determining the matching degree of the target user and the alternative package according to the weighted summation result of the matching distances between the target user and the alternative package in at least two package dimensions.
An electronic device 1500 according to this embodiment of the invention is described below with reference to fig. 15. The electronic device 1500 shown in fig. 15 is only an example and should not bring any limitation to the functions and the scope of use of the embodiments of the present invention.
As shown in fig. 15, electronic device 1500 is in the form of a general purpose computing device. Components of electronic device 1500 may include, but are not limited to: the at least one processing unit 1510, the at least one memory unit 1520, and the bus 1530 that connects the various system components (including the memory unit 1520 and the processing unit 1510).
Wherein the memory unit stores program code that is executable by the processing unit 1510 to cause the processing unit 1510 to perform steps according to various exemplary embodiments of the present invention as described in the above section "exemplary methods" of the present specification. For example, the processing unit 1510 may perform the steps as shown in fig. 2 or fig. 3 or fig. 4 or fig. 5 or fig. 6 or fig. 7 or fig. 13.
The storage unit 1520 may include readable media in the form of volatile storage units, such as a random access memory unit (RAM)15201 and/or a cache memory unit 15202, and may further include a read only memory unit (ROM) 15203.
Storage unit 1520 may also include a program/utility 15204 having a set (at least one) of program modules 15205, such program modules 15205 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may comprise an implementation of a network environment.
Bus 1530 may be any bus representing one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.
The electronic device 1500 can also communicate with one or more external devices 1600 (e.g., keyboard, pointing device, bluetooth device, etc.), with one or more devices that enable a user to interact with the electronic device 1500, and/or with any devices (e.g., router, modem, etc.) that enable the electronic device 1500 to communicate with one or more other computing devices. Such communication may occur via input/output (I/O) interface 1550. Also, the electronic device 1500 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network, such as the internet) via the network adapter 1560. As shown, the network adapter 1560 communicates with the other modules of the electronic device 1500 over the bus 1530. It should be appreciated that although not shown, other hardware and/or software modules may be used in conjunction with the electronic device 1500, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.
Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein may be implemented by software, or by software in combination with necessary hardware. Therefore, the technical solution according to the embodiments of the present disclosure may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (which may be a CD-ROM, a usb disk, a removable hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which may be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the embodiments of the present disclosure.
In an exemplary embodiment of the present disclosure, there is also provided a computer-readable storage medium having stored thereon a program product capable of implementing the above-described method of the present specification. In some possible embodiments, aspects of the invention may also be implemented in the form of a program product comprising program code means for causing a terminal device to carry out the steps according to various exemplary embodiments of the invention described in the above section "exemplary methods" of the present description, when said program product is run on the terminal device.
The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
A computer readable signal medium may include a propagated data signal with readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A readable signal medium may also be any readable medium that is not a readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.
Program code embodied on a readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
Program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., through the internet using an internet service provider).
Furthermore, the above-described figures are merely schematic illustrations of processes involved in methods according to exemplary embodiments of the invention, and are not intended to be limiting. It will be readily understood that the processes shown in the above figures are not intended to indicate or limit the chronological order of the processes. In addition, it is also readily understood that these processes may be performed synchronously or asynchronously, e.g., in multiple modules.
Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

Claims (10)

1. A package recommendation method is characterized by comprising the following steps:
acquiring historical package consumption of a target user;
determining historical package dimension usage of the target user under at least two package dimensions for the historical package usage of the target user;
determining predicted package dimension usage of the target user in each package dimension according to historical package dimension usage of the target user in the package dimension;
determining the matching degree of the target user and the alternative packages according to the predicted package dimension usage of the target user under at least two package dimensions;
and determining a recommended package according to the matching degree of the alternative packages so as to send the recommended package to the target user.
2. The method of claim 1, the package dimensions comprising two or more of:
tariff, traffic volume, voice volume, broadband cost and broadband rate.
3. The method of claim 2, wherein determining the predicted package dimension usage by the target user for each package dimension based on historical package dimension usage by the target user for that package dimension comprises:
if the package dimension comprises the tariff, determining the predicted tariff of the target user according to the historical tariff time sequence in the historical package usage;
if the package dimension comprises flow, determining the predicted flow consumption of the target user according to a historical flow consumption time sequence in the historical package consumption;
and if the package dimension comprises the voice consumption, determining the predicted voice consumption of the target user according to the historical voice consumption time sequence in the historical package consumption.
4. The method of claim 3, wherein determining the predicted tariff for the target subscriber based on a historical time series of tariffs in the historical package volume comprises:
performing stationarity detection and white noise detection on the historical tariff time sequence;
and if the stationarity detection result is a weak stationary sequence and the white noise detection result is a non-white noise sequence, processing the historical tariff time sequence by adopting an ARIMA time sequence prediction model to obtain the predicted tariff of the target user.
5. The method of claim 4, further comprising:
acquiring a training sample set, wherein the training sample set comprises a sample tariff time sequence;
processing the sample tariff time sequence through the ARIMA time sequence prediction model to obtain a sample prediction tariff;
and adjusting parameters of the ARIMA time sequence prediction model according to the sample tariff time sequence and the sample prediction tariff to obtain the trained ARIMA time sequence prediction model.
6. The method of claim 4, further comprising:
determining a switching user for switching to a recommended package;
integrating the historical package usage of the switching user into an update sample set;
and optimizing parameters of the ARIMA time series prediction model according to the updated sample set.
7. The method of claim 1, wherein determining a degree of matching of the target user with an alternative package based on predicted package dimension usage by the target user in at least two package dimensions comprises:
calculating an absolute value of a difference value between predicted package dimension usage of the target user in each package dimension and candidate package dimension usage of the candidate package in the package dimension, and obtaining an initial distance between the target user and the candidate package in each package dimension;
calculating the reciprocal of the sum of the initial distance between the target user and the alternative package in each package dimension and the distance compensation, and determining the matching distance between the target user and the alternative package in each package dimension;
and determining the matching degree of the target user and the alternative package according to the weighted summation result of the matching distances between the target user and the alternative package in at least two package dimensions.
8. A package recommendation device, comprising:
the historical data acquisition module is used for acquiring the historical package consumption of the target user;
a package dimension quantifying module, configured to determine, for the historical package usage of the target user, historical package dimension usage of the target user in at least two package dimensions;
a package dimension prediction module, configured to determine, according to historical package dimension usage of the target user in each package dimension, predicted package dimension usage of the target user in the package dimension;
the matching degree calculation module is used for determining the matching degree of the target user and the alternative packages according to the predicted package dimension usage of the target user under at least two package dimensions;
and the package recommending module can be used for determining a recommended package according to the matching degree of the alternative packages so as to send the recommended package to the target user.
9. An electronic device, comprising:
at least one processor;
storage means for storing at least one program;
when executed by the at least one processor, cause the at least one processor to implement the method of any one of claims 1-7.
10. A computer-readable medium, on which a computer program is stored, which program, when being executed by a processor, is adapted to carry out the method of any one of claims 1-7.
CN202111509717.5A 2021-12-10 2021-12-10 Package recommendation method, package recommendation device, electronic equipment and computer readable medium Active CN114186129B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111509717.5A CN114186129B (en) 2021-12-10 2021-12-10 Package recommendation method, package recommendation device, electronic equipment and computer readable medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111509717.5A CN114186129B (en) 2021-12-10 2021-12-10 Package recommendation method, package recommendation device, electronic equipment and computer readable medium

Publications (2)

Publication Number Publication Date
CN114186129A true CN114186129A (en) 2022-03-15
CN114186129B CN114186129B (en) 2023-05-26

Family

ID=80604448

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111509717.5A Active CN114186129B (en) 2021-12-10 2021-12-10 Package recommendation method, package recommendation device, electronic equipment and computer readable medium

Country Status (1)

Country Link
CN (1) CN114186129B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115049464A (en) * 2022-07-04 2022-09-13 中国联合网络通信集团有限公司 Operator package recommendation method and device and computer readable storage medium

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105825311A (en) * 2015-01-05 2016-08-03 中国移动通信集团湖南有限公司 Package determining method and package determining system
CN106095895A (en) * 2016-06-07 2016-11-09 百度在线网络技术（北京）有限公司 Information-pushing method and device
CN107070971A (en) * 2016-12-30 2017-08-18 北京瑞星信息技术股份有限公司 The recommendation method and device of telecommunication service
US20170329785A1 (en) * 2015-06-23 2017-11-16 Tencent Technology (Shenzhen) Company Limited Application recommendation method, server, and computer readable medium
WO2018000210A1 (en) * 2016-06-28 2018-01-04 深圳狗尾草智能科技有限公司 User portrait-based skill package recommendation device and method
CN107896153A (en) * 2017-11-15 2018-04-10 中国联合网络通信集团有限公司 A kind of flow package recommendation method and device based on mobile subscriber's internet behavior
CN109995837A (en) * 2018-01-02 2019-07-09 中国移动通信有限公司研究院 A kind of service package recommended method, device and server
CN110351098A (en) * 2018-04-08 2019-10-18 华为技术有限公司 Price previewing method and relevant device
CN110428244A (en) * 2019-06-14 2019-11-08 广州乐摇摇信息科技有限公司 Package recommendation method and device
CN113420211A (en) * 2021-06-22 2021-09-21 中国联合网络通信集团有限公司 Package recommendation method and device and electronic equipment

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105825311A (en) * 2015-01-05 2016-08-03 中国移动通信集团湖南有限公司 Package determining method and package determining system
US20170329785A1 (en) * 2015-06-23 2017-11-16 Tencent Technology (Shenzhen) Company Limited Application recommendation method, server, and computer readable medium
CN106095895A (en) * 2016-06-07 2016-11-09 百度在线网络技术（北京）有限公司 Information-pushing method and device
WO2018000210A1 (en) * 2016-06-28 2018-01-04 深圳狗尾草智能科技有限公司 User portrait-based skill package recommendation device and method
CN107070971A (en) * 2016-12-30 2017-08-18 北京瑞星信息技术股份有限公司 The recommendation method and device of telecommunication service
CN107896153A (en) * 2017-11-15 2018-04-10 中国联合网络通信集团有限公司 A kind of flow package recommendation method and device based on mobile subscriber's internet behavior
CN109995837A (en) * 2018-01-02 2019-07-09 中国移动通信有限公司研究院 A kind of service package recommended method, device and server
CN110351098A (en) * 2018-04-08 2019-10-18 华为技术有限公司 Price previewing method and relevant device
CN110428244A (en) * 2019-06-14 2019-11-08 广州乐摇摇信息科技有限公司 Package recommendation method and device
CN113420211A (en) * 2021-06-22 2021-09-21 中国联合网络通信集团有限公司 Package recommendation method and device and electronic equipment

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
JING DU,HAOCHEN XU,ZHIXIAO TU: "Personalized recommendation model of traffic package based on user consumption behavior", 2019 IEEE INTERNATIONAL CONFERENCE ON SIGNAL,INFORMATION AND DATA PROCESSING *
刘永等: "数据挖掘在电信套餐预演中的应用研究", 《计算机工程与设计》 *
王文: "基于深度学习的电信套餐推荐模型研究", 《中国优秀硕士学位论文全文数据库（信息科技辑）》 *
蔡衡: "推荐系统中的协同过滤算法研究", 《中国优秀硕士学位论文全文数据库（信息科技辑）》 *
马栋坤: "基于机器学习的电信行业用户的智能套餐匹配模型", 《中国优秀硕士学位论文全文数据库（基础科学辑）》 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115049464A (en) * 2022-07-04 2022-09-13 中国联合网络通信集团有限公司 Operator package recommendation method and device and computer readable storage medium

Also Published As

Publication number Publication date
CN114186129B (en) 2023-05-26

Similar Documents

Publication Publication Date Title
CN107423859B (en) Combined mold building method and system
US20190272553A1 (en) Predictive Modeling with Entity Representations Computed from Neural Network Models Simultaneously Trained on Multiple Tasks
US20110125700A1 (en) User model processing device
US10558987B2 (en) System identification framework
WO2010010654A1 (en) Usage estimation device
CN111026853B (en) Target problem determining method and device, server and customer service robot
CN113159355A (en) Data prediction method, data prediction device, logistics cargo quantity prediction method, medium and equipment
CN116541610B (en) Training method and device for recommendation model
KR101860457B1 (en) Method for analyzing weather affect and apparatus for executing the method
CN114186129B (en) Package recommendation method, package recommendation device, electronic equipment and computer readable medium
CN115935185A (en) Training method and device for recommendation model
CN113205189B (en) Method for training prediction model, prediction method and device
CN112116397A (en) User behavior characteristic real-time processing method and device, storage medium and electronic equipment
CN112906896A (en) Information processing method and device and computing equipment
CN110704614B (en) Information processing method and device for predicting user group type in application
CN114817716A (en) Method, device, equipment and medium for predicting user conversion behaviors and training model
CN111178987B (en) Method and device for training user behavior prediction model
CN114549032A (en) Stock processing method and device, computer storage medium and electronic equipment
CN113379436A (en) Information processing method, device, computing equipment and medium
CN113743636B (en) Target operation prediction method and device, electronic equipment and storage medium
EP3200141A1 (en) Method for managing device replacements
CN116823407B (en) Product information pushing method, device, electronic equipment and computer readable medium
CN109218411A (en) Data processing method and device, computer readable storage medium, electronic equipment
CN116629639B (en) Evaluation information determining method and device, medium and electronic equipment
CN113554460B (en) Potential user identification method and device

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