Method, apparatus, device and medium for determining conversion probability of users to be converted to 5G

Abstract

The present invention provides a method, apparatus, device and medium for determining the conversion probability of 5G users to be converted. The method includes: obtaining non-5G user information for which the 5G conversion probability is to be determined; inputting the non-5G user information into a conversion probability determination model trained to convergence to generate the 5G conversion probability corresponding to the non-5G user information; and outputting the 5G conversion probability. In the method for determining the conversion probability of 5G users to be converted according to the present invention, by inputting the non-5G user information into a conversion probability determination model trained to convergence, the 5G conversion probability corresponding to the non-5G user information can be generated, and thus the 5G conversion probability is output. Compared with the manual determination method, the efficiency of determining 5G users to be converted is higher.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a method, apparatus, device, and medium for determining the conversion probability of users to be converted to 5G. Background Art

[0002] With the continuous development of communication network technology, the fifth-generation mobile communication technology (English full name: 5th Generation Mobile Communication Technology, abbreviated as: 5G) is about to replace the fourth-generation mobile communication technology (English full name: 4th Generation Mobile Communication Technology, abbreviated as: 4G) as the main communication network in use. Currently, 4G is still the main communication network in use. During the process of 5G gradually replacing 4G, if it is possible to determine which users can be converted to the 5G network, the efficiency of 5G conversion can be improved.

[0003] Currently, it is mainly through manual means to determine which users can be converted to the 5G network, including directly communicating with users, querying the cell to which the user terminal belongs, etc. This method of manually determining users to be converted to 5G has low efficiency. Summary of the Invention

[0004] The present invention provides a method, apparatus, device, and medium for determining the conversion probability of users to be converted to 5G, so as to solve the problem of low efficiency in currently determining users to be converted to 5G through manual means.

[0005] In a first aspect of the present invention, a method for determining the conversion probability of users to be converted to 5G is provided, including:

[0006] Obtaining non-5G user information for which the 5G conversion probability is to be determined;

[0007] Inputting the non-5G user information into a converged conversion probability determination model to generate a 5G conversion probability corresponding to the non-5G user information;

[0008] Outputting the 5G conversion probability.

[0009] Further, in the method as described above, the obtaining non-5G user information for which the 5G conversion probability is to be determined includes:

[0010] Obtaining the highest access network type of each user terminal within a preset time period from a signaling collection system;

[0011] Determining whether the highest access network type of each user terminal is 5G;

[0012] If it is determined that the highest access network type of each of the user terminals is not 5G, it is determined that the user terminal matches a non-5G user;

[0013] Obtain non-5G user information corresponding to each non-5G user.

[0014] Further, in the method described above, the conversion probability determination model includes a random forest classification model and an XGBoost classification model;

[0015] Inputting the non-5G user information into the conversion probability determination model trained to convergence to generate the 5G conversion probability corresponding to the non-5G user information includes:

[0016] Input the non-5G user information into the random forest classification model to output a corresponding first conversion probability;

[0017] Input the non-5G user information into the XGBoost classification model to output a corresponding second conversion probability;

[0018] Determine the 5G conversion probability corresponding to the non-5G user information according to the first conversion probability, the second conversion probability, and each preset probability threshold that matches.

[0019] Further, in the method described above, determining the 5G conversion probability corresponding to the non-5G user information according to the first conversion probability, the second conversion probability, and each preset probability threshold that matches includes:

[0020] Determine a first difference between the first conversion probability and the preset probability threshold that matches;

[0021] Determine a second difference between the second conversion probability and the preset probability threshold that matches;

[0022] If the first difference is greater than zero and the second difference is greater than zero, determine the larger of the first conversion probability and the second conversion probability as the 5G conversion probability corresponding to the non-5G user information;

[0023] If the first difference is less than or equal to zero and / or the second difference is less than or equal to zero, determine the smaller of the first conversion probability and the second conversion probability as the 5G conversion probability corresponding to the non-5G user information.

[0024] Further, in the method described above, before inputting the non-5G user information into the conversion probability determination model trained to convergence, it further includes:

[0025] Obtain training samples, where the training samples include: 5G user information newly converted within a historical time period and non-5G user information that has not been converted; the newly converted 5G user information is the user information corresponding to non-5G users converted to 5G users within the historical time period; the non-converted non-5G user information is the user information that has always been a non-5G user within the historical time period;

[0026] Input the training samples into a preset conversion probability determination model to train the preset conversion probability determination model;

[0027] Determine whether both the F1 score index and the AUC evaluation index of the preset conversion probability determination model reach the highest values;

[0028] If it is determined that both the F1 score index and the AUC evaluation index reach the highest values, then determine that the preset conversion probability determination model meets the convergence condition;

[0029] Determine the preset conversion probability determination model that meets the convergence condition as the conversion probability determination model trained to convergence.

[0030] Further, in the method as described above, the obtaining of the training samples includes:

[0031] Obtain non-5G user information and 5G user information at the initial moment in the historical time period;

[0032] Obtain non-5G user information and 5G user information at the end moment in the historical time period;

[0033] Determine the training samples according to the non-5G user information and 5G user information at the initial moment and the non-5G user information and 5G user information at the end moment.

[0034] Further, in the method as described above, the determining of the training samples according to the non-5G user information and 5G user information at the initial moment and the non-5G user information and 5G user information at the end moment includes:

[0035] Determine a part of the non-5G user information at the end moment as the non-converted non-5G user information within the historical time period according to a preset ratio;

[0036] Determine the different 5G user information between the 5G user information at the end moment and the 5G user information at the initial moment as the newly converted 5G user information.

[0037] A second aspect of the present invention provides a conversion probability determination device for 5G users to be converted, including:

[0038] An obtaining module, configured to obtain non-5G user information for which the 5G conversion probability is to be determined;

[0039] A generation module, configured to input the non-5G user information into a conversion probability determination model trained to convergence to generate a 5G conversion probability corresponding to the non-5G user information;

[0040] An output module, configured to output the 5G conversion probability.

[0041] Further, for the device as described above, the obtaining module is specifically configured to:

[0042] Obtain the highest access network type of each user terminal within a preset time period from a signaling acquisition system; determine whether the highest access network type of each user terminal is 5G; if it is determined that the highest access network type of each user terminal is not 5G, then determine that the user terminal matches a non-5G user; obtain non-5G user information corresponding to each non-5G user.

[0043] Further, for the device as described above, the conversion probability determination model includes a random forest classification model and an XGBoost classification model;

[0044] The generation module is specifically configured to:

[0045] Input the non-5G user information into the random forest classification model to output a corresponding first conversion probability; input the non-5G user information into the XGBoost classification model to output a corresponding second conversion probability; determine the 5G conversion probability corresponding to the non-5G user information according to the first conversion probability, the second conversion probability, and each preset probability threshold that matches.

[0046] Further, for the device as described above, when the generation module determines the 5G conversion probability corresponding to the non-5G user information according to the first conversion probability, the second conversion probability, and each preset probability threshold that matches, it is specifically configured to:

[0047] Determine a first difference between the first conversion probability and the preset probability threshold that matches; determine a second difference between the second conversion probability and the preset probability threshold that matches; if the first difference is greater than zero and the second difference is greater than zero, then determine the larger one of the first conversion probability and the second conversion probability as the 5G conversion probability corresponding to the non-5G user information; if the first difference is less than or equal to zero and / or the second difference is less than or equal to zero, then determine the smaller one of the first conversion probability and the second conversion probability as the 5G conversion probability corresponding to the non-5G user information.

[0048] Further, for the device as described above, the device further includes:

[0049] A training module, configured to obtain training samples, where the training samples include: 5G user information newly converted within a historical time period and non-5G user information that has not been converted; the newly converted 5G user information is the user information corresponding to non-5G users converted to 5G users within the historical time period; the non-converted non-5G user information is the user information of users who have always been non-5G users within the historical time period; input the training samples into a preset conversion probability determination model to train the preset conversion probability determination model; determine whether both the F1 score index and the AUC evaluation index of the preset conversion probability determination model reach the highest values; if it is determined that both the F1 score index and the AUC evaluation index reach the highest values, determine that the preset conversion probability determination model meets the convergence condition; determine the preset conversion probability determination model that meets the convergence condition as the conversion probability determination model trained to convergence.

[0050] Further, for the apparatus as described above, when the training module obtains training samples, it is specifically configured to:

[0051] Obtain non-5G user information and 5G user information at the initial moment within the historical time period; obtain non-5G user information and 5G user information at the final moment within the historical time period; determine training samples based on the non-5G user information and 5G user information at the initial moment and the non-5G user information and 5G user information at the final moment.

[0052] Further, for the apparatus as described above, when the training module determines training samples based on the non-5G user information and 5G user information at the initial moment and the non-5G user information and 5G user information at the final moment, it is specifically configured to:

[0053] Determine a part of the non-5G user information at the final moment as the non-converted non-5G user information within the historical time period according to a preset ratio; determine the different 5G user information between the 5G user information at the final moment and the 5G user information at the initial moment as the newly converted 5G user information.

[0054] A third aspect of the present invention provides an electronic device, including: a memory, a processor;

[0055] The memory; a memory for storing executable instructions of the processor;

[0056] Wherein, the processor is configured to execute the conversion probability determination method for the 5G user to be converted according to any one of the first aspect by the processor.

[0057] In a fourth aspect of the present invention, there is provided a computer-readable storage medium storing computer-executable instructions, which are used to implement the conversion probability determination method for the to-be-converted 5G user according to any one of the first aspect when executed by a processor.

[0058] In a fifth aspect of the present invention, there is provided a computer program product including a computer program, which implements the conversion probability determination method for the to-be-converted 5G user according to any one of the first aspect when executed by a processor.

[0059] The present invention provides a method, device, equipment and medium for determining the conversion probability of a to-be-converted 5G user. The method includes: obtaining non-5G user information for which the 5G conversion probability is to be determined; inputting the non-5G user information into a converged conversion probability determination model to generate a 5G conversion probability corresponding to the non-5G user information; and outputting the 5G conversion probability. The method for determining the conversion probability of a to-be-converted 5G user according to the present invention can generate a 5G conversion probability corresponding to non-5G user information by inputting the non-5G user information into a converged conversion probability determination model, and thus output the 5G conversion probability. Compared with the manual determination method, the efficiency of determining the to-be-converted 5G user is higher. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] The drawings herein are incorporated into the specification and form a part of the specification, showing embodiments consistent with the present invention and used together with the specification to explain the principles of the present invention.

[0061] Figure 1 It is a scenario diagram for implementing the method for determining the conversion probability of a to-be-converted 5G user according to an embodiment of the present invention;

[0062] Figure 2 It is a schematic flowchart of the method for determining the conversion probability of a to-be-converted 5G user according to the first embodiment of the present invention;

[0063] Figure 3 It is a schematic flowchart of the method for determining the conversion probability of a to-be-converted 5G user according to the second embodiment of the present invention;

[0064] Figure 4 It is a schematic overall flowchart architecture diagram of the method for determining the conversion probability of a to-be-converted 5G user according to the second embodiment of the present invention;

[0065] Figure 5 It is a schematic structural diagram of the device for determining the conversion probability of a to-be-converted 5G user according to the third embodiment of the present invention;

[0066] Figure 6 It is a schematic structural diagram of the device for determining the conversion probability of a to-be-converted 5G user according to the fourth embodiment of the present invention;

[0067] Figure 7 Schematic structural diagram of the electronic device provided in the fifth embodiment of the present invention.

[0068] Through the above-mentioned drawings, specific embodiments of the present invention have been shown, and more detailed descriptions will be given hereinafter. These drawings and written descriptions are not intended to limit the scope of the inventive concept in any way, but to illustrate the concept of the present invention to those skilled in the art by referring to specific embodiments. Detailed implementation manners

[0069] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the present invention. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0070] Hereinafter, the technical solutions of the present invention will be described in detail with specific embodiments. These several specific embodiments may be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. Hereinafter, the embodiments of the present invention will be described with reference to the drawings.

[0071] To clearly understand the technical solutions of the present application, the solutions of the prior art will be introduced in detail first. The fifth-generation mobile communication technology, abbreviated as 5G, is a new generation of broadband mobile communication technology with the characteristics of high speed, low latency, and large connection. It is a network infrastructure for realizing the interconnection of humans, machines, and things. The International Telecommunication Union has defined three categories of application scenarios for 5G, namely enhanced mobile broadband, ultra-reliable low-latency communication, and massive machine-type communication. Therefore, it is currently a trend to gradually enable users to use 5G. When gradually enabling users to switch from the currently used network to 5G, if it is possible to determine which users can be switched to the 5G network, the efficiency of 5G conversion can be improved.

[0072] Currently, it is mainly determined manually which users can be switched from 4G to the 5G network, including directly communicating with users, querying the cells to which the user terminals belong, etc. The efficiency of determining the 5G users to be switched in this manual way is relatively low.

[0073] Therefore, aiming at the problem of relatively low efficiency of manually determining the 5G users to be switched in the prior art, the inventors found in the research that to solve this problem, machine learning can be combined to automatically determine the 5G users to be switched, thereby improving the determination efficiency.

[0074] Specifically, first, obtain the non-5G user information for which the 5G conversion probability is to be determined. Input the non-5G user information into the conversion probability determination model trained to convergence to generate the 5G conversion probability corresponding to the non-5G user information, and finally output the 5G conversion probability. The conversion probability determination method for the to-be-converted 5G users of the present invention can generate the 5G conversion probability corresponding to the non-5G user information by inputting the non-5G user information into the conversion probability determination model trained to convergence, thereby outputting the 5G conversion probability. Compared with the manual determination method, the efficiency of determining the to-be-converted 5G users is higher.

[0075] Based on the above creative findings, the inventors proposed the technical solution of this application.

[0076] Next, the application scenario of the conversion probability determination method for the to-be-converted 5G users provided by the embodiments of the present invention will be introduced. As Figure 1 shown, where 1 is the first electronic device and 2 is the second electronic device. The network architecture of the application scenario corresponding to the conversion probability determination method for the to-be-converted 5G users provided by the embodiments of the present invention includes: the first electronic device 1 and the second electronic device 2. The second electronic device 2 stores non-5G user information.

[0077] When it is necessary to determine the conversion probability of the to-be-converted 5G users, the first electronic device 1 obtains the non-5G user information stored in the second electronic device 2. Then, the first electronic device 1 inputs the non-5G user information into the conversion probability determination model trained to convergence to generate the 5G conversion probability corresponding to the non-5G user information. Finally, the first electronic device 1 outputs the 5G conversion probability to a display device or a user terminal, etc., so that the user can view and perform subsequent processing procedures based on the 5G conversion probability.

[0078] Next, the embodiments of the present invention will be introduced with reference to the accompanying drawings of the specification.

[0079] Figure 2 is a flowchart of the conversion probability determination method for the to-be-converted 5G users provided by the first embodiment of the present invention. As Figure 2 shown, in this embodiment, the execution subject of the embodiment of the present invention is a conversion probability determination device for the to-be-converted 5G users, and this conversion probability determination device for the to-be-converted 5G users can be integrated in an electronic device. Then, the conversion probability determination method for the to-be-converted 5G users provided by this embodiment includes the following steps:

[0080] Step S101, obtain the non-5G user information for which the 5G conversion probability is to be determined.

[0081] In this embodiment, the non-5G user information for which the 5G conversion probability is to be determined can be at a certain moment. Meanwhile, the acquisition method can be to obtain it from a database storing non-5G user information, or to determine the non-5G user information for which the 5G conversion probability is to be determined based on the user information of the entire network.

[0082] Step S102: Input the non-5G user information into the conversion probability determination model trained to convergence to generate the 5G conversion probability corresponding to the non-5G user information.

[0083] In this embodiment, the conversion probability determination model has been pre-trained to a convergence state. The conversion probability determination model can include multiple classification models, such as a random forest classification model and an XGBoost classification model, and the 5G conversion probability corresponding to the non-5G user information is determined by synthesizing the results generated by multiple classification models.

[0084] Step S103: Output the 5G conversion probability.

[0085] In this embodiment, the 5G conversion probability can be output to the user terminal of the staff so that the staff can perform subsequent processing based on the determined 5G conversion probability, such as negotiating with the user on how to convert to 5G, and further determining whether to convert and the conversion time, etc.

[0086] A method for determining the conversion probability of a 5G user to be converted provided by an embodiment of the present invention includes: obtaining non-5G user information for which the 5G conversion probability is to be determined. Inputting the non-5G user information into the conversion probability determination model trained to convergence to generate the 5G conversion probability corresponding to the non-5G user information. Outputting the 5G conversion probability. In the method for determining the conversion probability of a 5G user to be converted according to the present invention, by inputting the non-5G user information into the conversion probability determination model trained to convergence, the 5G conversion probability corresponding to the non-5G user information can be generated, and thus the 5G conversion probability is output. Compared with the manual determination method, the efficiency of determining the 5G user to be converted is higher.

[0087] Figure 3 It is a schematic flowchart of the method for determining the conversion probability of a 5G user to be converted provided by the second embodiment of the present invention. As Figure 3 shown, the method for determining the conversion probability of a 5G user to be converted provided by this embodiment further refines each step on the basis of the method for determining the conversion probability of a 5G user to be converted provided by the previous embodiment of the present invention. Then the method for determining the conversion probability of a 5G user to be converted provided by this embodiment includes the following steps.

[0088] Step S201: Obtain the highest access network type of each user terminal within a preset time period from the signaling collection system.

[0089] In this embodiment, the signaling acquisition system may be a DPI signaling acquisition system, where DPI stands for Deep Packet Inspection (Chinese: Deep Packet Detection).

[0090] The preset time period can be set according to requirements, such as set to one week, one month, etc.

[0091] Determining the highest access network type is because the current number of 5G base stations cannot meet the needs of users. Even if a user uses a 5G user terminal, they may only be able to use the 4G network at a certain moment. Therefore, when the highest access network type of the user terminal is 5G within the preset time period, it can be determined that the user terminal matches 5G.

[0092] Step S202, determine whether the highest access network type of each user terminal is 5G.

[0093] In this embodiment, generally, the highest access network type of the user terminal represents the network type that the user terminal matches.

[0094] Step S203, if it is determined that the highest access network type of each user terminal is not 5G, then determine that the user terminal matches a non-5G user.

[0095] In this embodiment, since the highest access network type of the user terminal represents the network type that the user terminal matches, therefore, when the highest access network type of each user terminal is not 5G, it can be determined that the user terminal matches a non-5G user.

[0096] Step S204, obtain the non-5G user information corresponding to each non-5G user.

[0097] In this embodiment, in the database, data is generally stored in a way that matches users, user information, and user terminals. After determining the non-5G users, the corresponding non-5G user information can be determined according to the non-5G users.

[0098] The non-5G user information may include user basic attributes (gender, age, package, network age, development channel, user star level, user group, etc.), consumption ability (traffic, other service subscriptions, etc.), network stickiness (2 / 3 / 4 / 5G in-network duration, off-network duration, etc.), terminal attributes (terminal model, terminal type, price, etc.), communication ability (call duration, call times, received call duration, received call times, etc.), roaming attributes (international roaming times, provincial roaming times, etc.), location information (busy-hour frequent cell in the morning, busy-hour resident cell in the evening, top three cells with the longest stay, etc.).

[0099] It should be noted that the conversion probability determination model includes a random forest classification model and an XGBoost classification model.

[0100] Step S205: Input the non-5G user information into the random forest classification model to output the corresponding first conversion probability.

[0101] In this embodiment, the conversion probability determination model is composed of the fusion of a random forest classification model and an XGBoost classification model. Therefore, it is necessary to determine the first conversion probability output by the random forest classification model. This first conversion probability is the 5G conversion probability calculated by the random forest classification model.

[0102] Step S206: Input the non-5G user information into the XGBoost classification model to output the corresponding second conversion probability.

[0103] In this embodiment, the conversion probability determination model is composed of the fusion of a random forest classification model and an XGBoost classification model. Therefore, it is necessary to determine the second conversion probability output by the XGBoost classification model. This second conversion probability is the 5G conversion probability calculated by the XGBoost classification model.

[0104] In order to make the conversion probability determination model achieve better results, it can be trained before inputting the non-5G user information into the conversion probability determination model trained to convergence. The training process is as follows:

[0105] Obtain training samples, which include: the information of newly converted 5G users and the information of non-converted non-5G users within the historical time period. The information of newly converted 5G users is the user information corresponding to the conversion from non-5G users to 5G users within the historical time period. The information of non-converted non-5G users is the user information that has always been a non-5G user within the historical time period.

[0106] Input the training samples into the preset conversion probability determination model to train the preset conversion probability determination model.

[0107] Determine whether both the F1 score index and the AUC evaluation index of the preset conversion probability determination model reach the highest values.

[0108] If it is determined that both the F1 score index and the AUC evaluation index reach the highest values, it is determined that the preset conversion probability determination model meets the convergence condition.

[0109] Determine the preset conversion probability determination model that meets the convergence condition as the conversion probability determination model trained to convergence.

[0110] In this embodiment, when training the conversion probability determination model, training samples are required, and the training samples need to include the parameters input into the conversion probability determination model and the parameters for verifying whether the output of the conversion probability determination model is correct.

[0111] During training, the information of newly converted 5G users and non-5G users that have not been converted within the historical time period can be trained separately according to different moments in the historical time period. At the same time, the information of newly converted 5G users and non-5G users that have not been converted at a certain moment within the historical time period can be marked to identify which non-5G terminal users have been converted into 5G terminal users. Those who have been converted from non-5G terminal users to 5G terminal users within the statistical period are positive examples of our target and are marked as 1, while those who are still non-5G terminal users are negative examples of our target and are marked as 0. Tagging these two types of users is the positive example set and the negative example set.

[0112] The F1-score (fully known as F1-score in English) is a measurement index for classification problems. In some machine learning competitions for multi-classification problems, the F1-score is often used as the final evaluation method. It is the harmonic mean of precision and recall, with a maximum value of 1 and a minimum value of 0. The AUC is a model evaluation index in the field of machine learning. The AUC (fully known as area under the curve in English) is the area under the receiver operating characteristic curve, and the value of this area will not be greater than 1.

[0113] At the same time, the steps to obtain training samples can be specifically as follows:

[0114] Obtain the information of non-5G users and 5G users at the initial moment in the historical time period.

[0115] Obtain the information of non-5G users and 5G users at the final moment in the historical time period.

[0116] Determine the training samples based on the information of non-5G users and 5G users at the initial moment and the information of non-5G users and 5G users at the final moment.

[0117] Since what needs to be determined is the process of converting from non-5G to 5G, therefore, obtain the information of non-5G users and 5G users at the initial moment in the historical time period, and then compare it with the information of non-5G users and 5G users at the final moment in the historical time period to determine the training samples.

[0118] Optionally, in this embodiment, determining the training samples based on the information of non-5G users and 5G users at the initial moment and the information of non-5G users and 5G users at the final moment includes:

[0119] Determine a part of the information of non-5G users at the final moment as the information of non-5G users that have not been converted within the historical time period according to a preset ratio.

[0120] Determine the different 5G user information between the 5G user information at the final moment and the 5G user information at the initial moment as the information of newly converted 5G users.

[0121] Since it is necessary to determine the newly converted 5G user information, it is necessary to compare the 5G user information at the last moment with the 5G user information at the initial moment, so as to determine the information difference between the two. This information difference is the newly converted 5G user information.

[0122] Step S207: Determine the 5G conversion probability corresponding to the non-5G user information according to the first conversion probability, the second conversion probability, and each preset probability threshold that matches.

[0123] In this embodiment, both the first conversion probability and the second conversion probability have matching preset probability thresholds. These preset probability thresholds may be the same or different, and at the same time, they can also be set according to actual needs.

[0124] Optionally, in this embodiment, determining the 5G conversion probability corresponding to the non-5G user information according to the first conversion probability, the second conversion probability, and each preset probability threshold that matches includes:

[0125] Determine the first difference between the first conversion probability and the matching preset probability threshold.

[0126] Determine the second difference between the second conversion probability and the matching preset probability threshold.

[0127] If the first difference is greater than zero and the second difference is greater than zero, then determine the larger one of the first conversion probability and the second conversion probability as the 5G conversion probability corresponding to the non-5G user information.

[0128] If the first difference is less than or equal to zero and / or the second difference is less than or equal to zero, then determine the smaller one of the first conversion probability and the second conversion probability as the 5G conversion probability corresponding to the non-5G user information.

[0129] In this embodiment, when both differences, that is, the first difference and the second difference, are greater than zero, it means that the results output by the two models can meet the prediction requirements. Therefore, the larger one of the first conversion probability and the second conversion probability can be determined as the 5G conversion probability corresponding to the non-5G user information.

[0130] If there is a situation where the first difference and / or the second difference is less than zero, it means that there may be a large calculation error in one of the two models. For the sake of conservatism, the smaller one of the first conversion probability and the second conversion probability can be retained and determined as the 5G conversion probability corresponding to the non-5G user information.

[0131] Step S208: Output the 5G conversion probability.

[0132] In this embodiment, the implementation of step 208 is similar to the implementation of step 103 in the previous embodiment of the present invention, and will not be described in detail here.

[0133] In order to better understand the method for determining the conversion probability of a 5G user to be converted in this embodiment, the following will be described in detail using an actual application scenario as an example. Figure 4 As shown, this embodiment explains the process from data collection, data cleaning, feature engineering, building a conversion probability determination model, and determining the conversion probability of the 5G users to be converted.

[0134] Data collection: Based on a DPI signaling collection system, 5G terminal users and non-5G terminal users are identified according to the highest access network type and terminal TAC (Type Allocation Code) configuration library.

[0135] For all non-5G terminal users in the first week of a month, this list of users is associated with the data of the first week of the next month, and the numbers that have been converted to 5G terminal users are identified and labeled as 1, and the numbers of non-5G terminal users that have not been converted are labeled as 0. Those labeled 1 are positive examples, and those labeled 0 are negative examples.

[0136] For users with positive and negative example labels, user information is determined, mainly including basic user attributes (gender, age, package, network age, user star rating, user group, etc.), consumption capacity (traffic, other service subscriptions), network stickiness (2 / 3 / 4 / 5G network time, off-network time, etc.), terminal attributes (terminal manufacturer, terminal model, launch date, terminal standard, price, etc.), communication capabilities (calling duration, calling number, called duration, called number, etc.), roaming attributes (national roaming number, provincial roaming number, etc.), location information (the cell where you often stay during the morning busy period, the cell where you often stay during the evening busy period, the top three cells with the longest stay, etc.). When associating data, grouping is performed according to the number and terminal IMEI (English full name: International Mobile Equipment Identity, Chinese: International Mobile Equipment Identity Code), and one record is retained for one user. Since the proportion of negative example data is too high, we first use the undersampling method to sample the negative example data so that the ratio of positive examples to negative examples is about 1:5. In this way, the target data set is generated, and then the target data set is divided into training set and test set according to the ratio of 7:3.

[0137] Data cleaning and feature engineering: Fill in the null values in the above dataset. For numerical null values, we use zero filling or mean filling. For outliers, we eliminate or smooth them. For fields with a null value rate exceeding 60%, we directly eliminate them. For text fields, we perform category conversion, etc. Analyze the correlation features between fields through feature engineering on the cleaned data, then perform principal component analysis on some features, and retain several principal components with a contribution rate greater than 85%. After data cleaning and feature engineering, we obtain the training set and test set that can be used for modeling.

[0138] Construct a conversion probability determination model: Build a conversion probability determination model by fusing a random forest classification model and an XGBoost classification model. First, build a random forest classification model and use the Bayesian hyperparameter tuning method to optimize the parameters to maximize the F1 score of the random forest classification model. Subsequently, build an XGBoost classification model and use the random search method to optimize the model parameters to maximize the AUC of the model. At this point, both classification models have been established. For the non-5G user information for which the conversion probability needs to be determined, first, the random forest model is used for classification prediction, and then the XGBoost classification model is used for prediction. As Figure 4 shown, the RF classification model refers to the random forest model, p1 represents the prediction probability of the random forest model, a represents the preset probability threshold corresponding to the random forest model, p2 represents the prediction probability of the XGBoost classification model, and b represents the preset probability threshold corresponding to the XGBoost classification model. If the prediction results of both models are greater than the preset probability threshold, then the prediction result of the user is selected as the maximum value of the two probabilities; otherwise, the prediction result is taken as the minimum value of the two probabilities.

[0139] Figure 5 This is the structural schematic diagram of the conversion probability determination device for the to-be-converted 5G users provided in the third embodiment of the present invention. As Figure 5 shown, in this embodiment, the conversion probability determination device 300 for the to-be-converted 5G users includes:

[0140] An acquisition module 301, configured to acquire non-5G user information for which the 5G conversion probability needs to be determined.

[0141] A generation module 302, configured to input the non-5G user information into the converged conversion probability determination model to generate the 5G conversion probability corresponding to the non-5G user information.

[0142] An output module 303, configured to output the 5G conversion probability.

[0143] The conversion probability determination device for the to-be-converted 5G users provided in this embodiment can execute Figure 2 the technical solutions of the method embodiment shown, and its implementation principle and technical effects are the same as those of Figure 2The method embodiments are similar to those shown above and will not be elaborated one by one here.

[0144] Meanwhile, Figure 6 FIG. is a schematic structural diagram of a conversion probability determination device for a to-be-converted 5G user provided in the fourth embodiment of the present invention. As Figure 6 shown, based on the conversion probability determination device for a to-be-converted 5G user provided in the previous embodiment, the conversion probability determination device for a to-be-converted 5G user provided by the present invention is further refined. To distinguish it from the previous embodiment, the conversion probability determination device for a to-be-converted 5G user in this embodiment is described as the conversion probability determination device 400 for a to-be-converted 5G user.

[0145] Optionally, in this embodiment, the acquisition module 301 is specifically configured to:

[0146] Obtain the highest access network type of each user terminal within a preset time period from the signaling acquisition system. Determine whether the highest access network type of each user terminal is 5G. If it is determined that the highest access network type of each user terminal is not 5G, then determine that the user terminal matches a non-5G user. Obtain the non-5G user information corresponding to each non-5G user.

[0147] Optionally, in this embodiment, the conversion probability determination model includes a random forest classification model and an XGBoost classification model.

[0148] The generation module 302 is specifically configured to:

[0149] Input the non-5G user information into the random forest classification model to output the corresponding first conversion probability. Input the non-5G user information into the XGBoost classification model to output the corresponding second conversion probability. Determine the 5G conversion probability corresponding to the non-5G user information according to the first conversion probability, the second conversion probability, and each preset probability threshold that matches.

[0150] Optionally, in this embodiment, when the generation module 302 determines the 5G conversion probability corresponding to the non-5G user information according to the first conversion probability, the second conversion probability, and each preset probability threshold that matches, it is specifically configured to:

[0151] Determine the first difference between the first conversion probability and the preset probability threshold that matches. Determine the second difference between the second conversion probability and the preset probability threshold that matches. If the first difference is greater than zero and the second difference is greater than zero, then determine the larger value of the first conversion probability and the second conversion probability as the 5G conversion probability corresponding to the non-5G user information. If the first difference is less than or equal to zero and / or the second difference is less than or equal to zero, then determine the smaller value of the first conversion probability and the second conversion probability as the 5G conversion probability corresponding to the non-5G user information.

[0152] Optionally, in this embodiment, the conversion probability determination device 400 for the 5G users to be converted further includes:

[0153] A training module 401, configured to obtain training samples, where the training samples include: information of newly converted 5G users and information of non-5G users who have not been converted within a historical time period. The information of newly converted 5G users is the user information corresponding to non-5G users converted to 5G users within the historical time period. The information of non-5G users who have not been converted is the user information of non-5G users who have always been non-5G users within the historical time period. Input the training samples into a preset conversion probability determination model to train the preset conversion probability determination model. Determine whether both the F1 score index and the AUC evaluation index of the preset conversion probability determination model reach the maximum value. If it is determined that both the F1 score index and the AUC evaluation index reach the maximum value, it is determined that the preset conversion probability determination model meets the convergence condition. Determine the preset conversion probability determination model that meets the convergence condition as the conversion probability determination model trained to convergence.

[0154] Optionally, in this embodiment, when the training module 401 obtains training samples, it is specifically configured to:

[0155] Obtain the non-5G user information and 5G user information at the initial moment in the historical time period. Obtain the non-5G user information and 5G user information at the final moment in the historical time period. Determine the training samples according to the non-5G user information and 5G user information at the initial moment and the non-5G user information and 5G user information at the final moment.

[0156] Optionally, in this embodiment, when the training module 401 determines the training samples according to the non-5G user information and 5G user information at the initial moment and the non-5G user information and 5G user information at the final moment, it is specifically configured to:

[0157] Determine a part of the non-5G user information at the final moment as the non-5G user information that has not been converted within the historical time period according to a preset ratio. Determine the different 5G user information between the 5G user information at the final moment and the 5G user information at the initial moment as the information of newly converted 5G users.

[0158] The conversion probability determination device for the 5G users to be converted provided in this embodiment may execute Figures 2 - 4 the technical solution of the method embodiment shown, and its implementation principle and technical effect are similar to those of Figures 2 - 4 the method embodiment shown, and will not be elaborated here one by one.

[0159] According to the embodiments of the present invention, the present invention also provides an electronic device, a computer-readable storage medium, and a computer program product.

[0160] As Figure 7 shown, Figure 7It is a schematic structural diagram of an electronic device provided by the fifth embodiment of the present invention. The electronic device is intended for various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0161] As Figure 7 shown, the electronic device includes: a processor 501 and a memory 502. Each component is interconnected using different buses and can be mounted on a common motherboard or otherwise installed as needed. The processor can process instructions executed within the electronic device.

[0162] The memory 502 is the non-transitory computer-readable storage medium provided by the present invention. Among them, the memory stores instructions executable by at least one processor, so that at least one processor executes the method for determining the conversion probability of the to-be-converted 5G user provided by the present invention. The non-transitory computer-readable storage medium of the present invention stores computer instructions, and the computer instructions are used to cause a computer to execute the method for determining the conversion probability of the to-be-converted 5G user provided by the present invention.

[0163] The memory 502, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the program instructions / modules corresponding to the method for determining the conversion probability of the to-be-converted 5G user in the embodiment of the present invention (for example, the Figure 5 acquisition module 301, generation module 302, and output module 303 shown). The processor 501 executes various functional applications and data processing of the server by running the non-transitory software programs, instructions, and modules stored in the memory 502, that is, implements the method for determining the conversion probability of the to-be-converted 5G user in the above method embodiments.

[0164] Meanwhile, this embodiment also provides a computer product. When the instructions in the computer product are executed by the processor of the electronic device, the electronic device can execute the method for determining the conversion probability of the to-be-converted 5G user in the above-mentioned first to second embodiments.

[0165] Other embodiments of the embodiments of the present invention will be readily apparent to those skilled in the art after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses, or adaptations of the embodiments of the present invention, which follow the general principles of the embodiments of the present invention and include known common knowledge or conventional technical means in the technical field not disclosed in the embodiments of the present invention. The specification and examples are only to be considered as exemplary, and the true scope and spirit of the embodiments of the present invention are pointed out by the following claims.

[0166] It should be understood that the embodiments of the present invention are not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the embodiments of the present invention is only limited by the appended claims.

Claims

1. A method for determining the conversion probability of a 5G user to be converted, characterized in that, Including: Obtain non-5G user information for which the 5G conversion probability is to be determined; Input the non-5G user information into a conversion probability determination model trained to convergence to generate the 5G conversion probability corresponding to the non-5G user information; Output the 5G conversion probability; The conversion probability determination model includes a random forest classification model and an XGBoost classification model; The step of inputting the non-5G user information into a conversion probability determination model trained to convergence to generate the 5G conversion probability corresponding to the non-5G user information includes: Input the non-5G user information into the random forest classification model to output a corresponding first conversion probability; Input the non-5G user information into the XGBoost classification model to output a corresponding second conversion probability; Determine a first difference between the first conversion probability and a matching preset probability threshold; Determine a second difference between the second conversion probability and a matching preset probability threshold; If the first difference is greater than zero and the second difference is greater than zero, then determine the larger one of the first conversion probability and the second conversion probability as the 5G conversion probability corresponding to the non-5G user information; If the first difference is less than or equal to zero and / or the second difference is less than or equal to zero, then determine the smaller one of the first conversion probability and the second conversion probability as the 5G conversion probability corresponding to the non-5G user information.

2. The method according to claim 1, wherein The step of obtaining non-5G user information for which the 5G conversion probability is to be determined includes: Obtain the highest access network type of each user terminal within a preset time period from a signaling collection system; Determine whether the highest access network type of each user terminal is 5G; If it is determined that the highest access network type of each user terminal is not 5G, then determine that the user terminal matches a non-5G user; Obtain the non-5G user information corresponding to each non-5G user.

3. The method according to claim 1 or 2, characterized in that, Before inputting the non-5G user information into a conversion probability determination model trained to convergence, it further includes: Obtain training samples, where the training samples include: 5G user information newly converted within a historical time period and non-5G user information that has not been converted; The newly converted 5G user information is the user information corresponding to a non-5G user converted to a 5G user within a historical time period; The non-5G user information that has not been converted is the user information of a user who has always been a non-5G user within a historical time period; Input the training samples into a preset conversion probability determination model to train the preset conversion probability determination model; Determine whether both the F1 score index and the AUC evaluation index of the preset conversion probability determination model reach the highest value; If it is determined that both the F1 score index and the AUC evaluation index reach the highest value, then determine that the preset conversion probability determination model meets the convergence condition; Determine the preset conversion probability determination model that meets the convergence condition as the conversion probability determination model trained to convergence.

4. The method according to claim 3, characterized in that, The step of obtaining training samples includes: Obtain the non-5G user information and 5G user information at the initial moment within a historical time period; Obtain the non-5G user information and 5G user information at the final moment within a historical time period; Determine training samples based on the non-5G user information and 5G user information at the initial moment, and the non-5G user information and 5G user information at the final moment.

5. The method according to claim 4, wherein The determination of training samples based on the non-5G user information and 5G user information at the initial moment, and the non-5G user information and 5G user information at the final moment includes: Determine a part of the non-5G user information at the final moment as the non-5G user information that has not been converted within the historical time period according to a preset ratio; Determine the different 5G user information between the 5G user information at the final moment and the 5G user information at the initial moment as the newly converted 5G user information.

6. A conversion probability determination device for a 5G user to be converted, characterized in that, Includes: An acquisition module for acquiring non-5G user information for which the 5G conversion probability is to be determined; A generation module for inputting the non-5G user information into a conversion probability determination model trained to convergence to generate the 5G conversion probability corresponding to the non-5G user information; An output module for outputting the 5G conversion probability; The conversion probability determination model includes a random forest classification model and an XGBoost classification model; the generation module is specifically used for: Input the non-5G user information into the random forest classification model to output the corresponding first conversion probability; Input the non-5G user information into the XGBoost classification model to output the corresponding second conversion probability; Determine the first difference between the first conversion probability and the matching preset probability threshold; Determine the second difference between the second conversion probability and the matching preset probability threshold; If the first difference is greater than zero and the second difference is greater than zero, determine the larger one of the first conversion probability and the second conversion probability as the 5G conversion probability corresponding to the non-5G user information; If the first difference is less than or equal to zero and / or the second difference is less than or equal to zero, determine the smaller one of the first conversion probability and the second conversion probability as the 5G conversion probability corresponding to the non-5G user information.

7. An electronic device, characterized in that, Includes: A processor and a memory communicatively connected to the processor; The memory stores computer execution instructions; The processor executes the computer execution instructions stored in the memory to implement the method for determining the conversion probability of a 5G user to be converted as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, Computer execution instructions are stored in the computer-readable storage medium, and when the computer execution instructions are executed by a processor, they are used to implement the method for determining the conversion probability of a 5G user to be converted as described in any one of claims 1 to 5.

Images