Promotion information processing method and device and electronic equipment

Abstract

The invention discloses a promotion information processing method and device and electronic equipment. Relates to the big data field. The method comprises the following steps: acquiring passenger information of a target passenger, and carrying out numerical processing on the passenger information to obtain passenger data; the passenger type of the target passenger is determined according to the passenger information, a probability prediction model is selected according to the passenger type, a target probability prediction model is obtained, and the target probability prediction model is used for predicting the probability that the target passenger handles the lift-up service; and inputting the passenger data into the target probability prediction model to obtain a prediction result of the probability that the target passenger handles the lift-up service, and determining a processing result of promotion information according to the prediction result, the promotion information being used for promoting the lift-up service to the target passenger. Through the method and the device, the problem of relatively low promotion efficiency of the cabin lifting service in related technologies is solved.

Description

Technical Field

[0001] The present application relates to the field of big data, and more specifically, to a method, device and electronic device for processing promotion information. Background Art

[0002] In recent years, the civil aviation industry has developed rapidly, and all airlines are facing fierce market competition. Accurate control of the probability of passenger upgrades can help airlines optimize pricing strategies and ensure the best use of seats, thereby increasing revenue, and can also help airlines better meet customer expectations. If a passenger expects an upgrade and eventually gets the upgrade opportunity, they may have a higher degree of satisfaction with the airline's services, which helps retain and attract customers.

[0003] When promoting the upgrade service to users, the upgrade service information is usually pushed to all users indiscriminately, which will make a large amount of pushed information invalid and increase the push cost, or the upgrade service information is only pushed to users who have performed upgrade operations multiple times. However, this will greatly reduce the efficiency of mining potential upgrade users.

[0004] With regard to the problem of low promotion efficiency of cabin upgrade services in related technologies, no effective solution has been proposed so far. Summary of the invention

[0005] The present application provides a promotion information processing method, device and electronic device to solve the problem of low promotion efficiency of upgrade services in related technologies.

[0006] According to one aspect of the present application, a method for processing promotion information is provided. The method includes: obtaining passenger information of a target passenger, and numerically processing the passenger information to obtain passenger data; determining the passenger type of the target passenger according to the passenger information, and selecting a probability prediction model according to the passenger type to obtain a target probability prediction model, wherein the target probability prediction model is used to predict the probability of the target passenger applying for an upgrade; inputting the passenger data into the target probability prediction model to obtain a prediction result of the probability of the target passenger applying for an upgrade, and determining a processing result of the promotion information according to the prediction result, wherein the promotion information is used to promote the upgrade to the target passenger.

[0007] Optionally, the passenger information includes M characteristic values, and the passenger information of the target passenger is obtained, and the passenger information is numerically processed to obtain passenger data, including: obtaining non-numerical characteristic values ​​in the passenger information to obtain N first characteristic values, wherein M and N are positive integers, and M is equal to N; obtaining the numerical type of each first characteristic value, and processing each first characteristic value according to the processing method corresponding to the numerical type to obtain N updated characteristic values, wherein each updated characteristic value is a numerical value; combining the N updated characteristic values ​​with the numerical characteristic values ​​in the passenger information to obtain passenger data.

[0008] Optionally, each first eigenvalue is processed according to a processing method corresponding to the numerical type to obtain N updated eigenvalues, including: when the numerical type of the first eigenvalue is the first type, obtaining a comparison table of the first eigenvalue, and determining the level of the eigenvalue according to the comparison table, wherein the first type represents that the eigenvalue has a corresponding level, and the comparison table contains multiple eigenvalue intervals and the level of each eigenvalue interval; using the level to replace the first eigenvalue to obtain an updated eigenvalue; when the numerical type of the first eigenvalue is the second type, encoding the first eigenvalue by a preset encoding method to obtain an updated eigenvalue, wherein the second type represents that the eigenvalue does not have a corresponding level.

[0009] Optionally, selecting a probability prediction model according to the passenger type to obtain a target probability prediction model includes: obtaining the frequency of target passengers taking flights in a historical time period in a database to obtain a target frequency; determining the frequency interval to which the target frequency belongs, and determining the probability prediction model associated with the frequency interval as the target probability prediction model, wherein each probability prediction model is associated with a different frequency interval.

[0010] Optionally, the target probability prediction model is trained in the following manner: historical passengers of the same passenger type as the target passenger in each institution are obtained to obtain P historical passenger sets, wherein each historical passenger set contains Q historical passengers, and P and Q are positive integers; historical passenger information and upgrade labels of each historical passenger in the P historical passenger sets are obtained to obtain P sample data sets, wherein each sample data set contains multiple sample data, each sample data contains historical passenger information and upgrade labels reserved for a historical passenger when taking a flight, and the upgrade label is determined according to whether the historical passenger selects the upgrade service each time he takes the flight; the initial prediction model is trained respectively with each sample data set to obtain P candidate prediction models; the model parameters in the initial prediction model are updated with the model parameters in the P candidate prediction models to obtain a first prediction model, and the training set of each candidate prediction model is updated, and the model parameters in the first prediction model are iteratively calculated by updating each candidate prediction model until the first prediction model converges to obtain the target probability prediction model.

[0011] Optionally, obtaining historical passenger information and upgrade labels of each historical passenger in P historical passenger sets to obtain P sample data sets includes: determining the historical passenger information and upgrade labels of each historical passenger each time he takes a flight as initial sample data to obtain Q initial sample data; judging whether there are missing values ​​in each initial sample data, and if there are missing values ​​in the target initial sample data, obtaining the number of missing values ​​to obtain a target number; if the target number is greater than a preset number, deleting the target initial sample data; if the target number is less than or equal to the preset number, filling the missing values ​​to obtain updated target initial sample data; after processing each of the Q initial sample data, taking the Q updated initial sample data as a sample data set to obtain P sample data sets.

[0012] Optionally, filling missing values ​​to obtain updated target initial sample data includes: when the missing value is categorical, replacing the missing value with a preset value; when the missing value is numerical, obtaining the target feature value corresponding to the missing value in the initial sample data with the highest similarity to the target initial sample data, and replacing the missing value with the target feature value.

[0013] Optionally, determining the processing result of the promotion information based on the prediction result includes: when the prediction result indicates that the probability of the target passenger's upgrade is greater than or equal to a preset probability, executing the operation of sending the promotion information to the target passenger; when the prediction result indicates that the probability of the target passenger's upgrade is less than a preset probability, not executing the operation of sending the promotion information to the target passenger.

[0014] According to another aspect of the present application, a device for processing promotion information is provided. The device includes: a first acquisition unit, which is used to acquire passenger information of a target passenger, and numerically process the passenger information to obtain passenger data; a determination unit, which is used to determine the passenger type of the target passenger based on the passenger information, and select a probability prediction model based on the passenger type to obtain a target probability prediction model, wherein the target probability prediction model is used to predict the probability of the target passenger applying for an upgrade; a prediction unit, which is used to input the passenger data into the target probability prediction model, obtain a prediction result of the probability of the target passenger applying for an upgrade, and determine a processing result of the promotion information based on the prediction result, wherein the promotion information is used to promote the upgrade service to the target passenger.

[0015] According to another aspect of the present invention, a computer program product is provided, including a computer program, which, when executed by a processor, implements a method for processing promotion information provided in the aforementioned embodiment of the present application.

[0016] According to another aspect of the present invention, an electronic device is provided, comprising one or more processors and a memory; the memory stores computer-readable instructions, and the processor is used to execute the computer-readable instructions, wherein when the computer-readable instructions are executed, a method for processing promotion information provided in the aforementioned embodiment is executed.

[0017] Through this application, the following steps are adopted: obtaining the passenger information of the target passenger, and processing the passenger information numerically to obtain passenger data; determining the passenger type of the target passenger according to the passenger information, and selecting a probability prediction model according to the passenger type to obtain a target probability prediction model, wherein the target probability prediction model is used to predict the probability of the target passenger applying for an upgrade service; inputting the passenger data into the target probability prediction model to obtain a prediction result of the probability of the target passenger applying for an upgrade service, and determining the processing result of the promotion information according to the prediction result, wherein the promotion information is used to promote the upgrade service to the target passenger. The problem of low promotion efficiency of the upgrade service in the related technology is solved. By processing the passenger information and determining the passenger type according to the passenger information, a model for predicting the probability of upgrading the passenger is determined, and the target passenger's upgrade probability is predicted by the target probability prediction model, and then determining whether to promote the upgrade service to the user according to the upgrade probability, thereby achieving the technical effect of improving the promotion efficiency of the upgrade service. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0019] Figure 1 is a flowchart of a method for processing promotion information provided in an embodiment of the present application;

[0020] Figure 2 It is a flow chart for performing numerical processing of data according to an embodiment of the present application;

[0021] Figure 3 is a schematic diagram of a promotion information processing device provided according to an embodiment of the present application;

[0022] Figure 4 It is a schematic diagram of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0023] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0024] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0025] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0026] It should be noted that the promotion information processing method, device and electronic device determined in the present disclosure can be used in the big data field, and can also be used in any field other than the big data field. The application field of the promotion information processing method, device and electronic device determined in the present disclosure is not limited.

[0027] It should be noted that the collected information, user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) used in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data are in compliance with the relevant laws, regulations and standards of the relevant regions, necessary confidentiality measures are taken, and public order and good customs are not violated. Corresponding operation portals are provided for users to choose to authorize or refuse to use. For example, an interface is set up between this system and relevant users or organizations. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or organization through the interface, and obtain relevant information after receiving the consent information fed back by the aforementioned user or organization.

[0028] The embodiments or examples of the present disclosure are not exhaustive, but are only illustrative of some embodiments or examples, and are not intended to be specific limitations on the scope of protection of the present disclosure. In the absence of contradiction, each step in a certain embodiment or example can be implemented as an independent example, and the steps can be combined arbitrarily. For example, the scheme after removing some steps in a certain embodiment or example can also be implemented as an independent example, and the order of the steps in a certain embodiment or example can be arbitrarily exchanged. In addition, the optional methods or optional examples in a certain embodiment or example can be combined arbitrarily; in addition, the various embodiments or examples can be combined arbitrarily, for example, some or all steps of different embodiments or examples can be combined arbitrarily, and a certain embodiment or example can be combined arbitrarily with the optional methods or optional examples of other embodiments or examples.

[0029] For the convenience of description, some nouns or terms involved in the embodiments of the present application are explained below:

[0030] Federated learning is an approach to distributed machine learning that allows multiple devices or data centers to collaboratively train models without sharing data.

[0031] Figure 1 FIG. 1 is a flowchart of a method for processing promotion information provided in an embodiment of the present application. Figure 1 As shown, the method comprises the following steps:

[0032] Step S101, obtaining the passenger information of the target passenger, and digitizing the passenger information to obtain passenger data.

[0033] Specifically, the passenger information of the target passenger may include the passenger information related to the flight that the passenger has reserved when taking the current flight, such as passenger personal information: whether the passenger is a major customer or a frequent flyer; passenger ticket information: travel date, booking time, ticket issuance time, flight fare, flight discount, cabin class, whether to voluntarily upgrade, whether to voluntarily downgrade, and paid upgrade amount; passenger itinerary information: whether the passenger is traveling in a group. It may also include flight information of the flight the passenger is taking, which may include flight information: departure time, arrival time, departure airport, arrival airport, flight duration, whether the passenger is traveling during the Spring Festival travel rush, whether the passenger is traveling on holidays, flight distance, and flight delay rate; other information: economic prosperity of the arrival airport and economic prosperity of the arrival airport.

[0034] Furthermore, after obtaining the passenger information, since the passenger information contains a lot of non-numerical information, in order to ensure that the prediction model can process the passenger information normally, it is necessary to first digitize the passenger information, so as to obtain passenger data in which all information is numerical information.

[0035] For example, after digitizing the passenger information, the following passenger data can be obtained: the time period of the passenger's flight's departure time; the time period of the passenger's flight's arrival time; the prosperity of the departure airport of the passenger's flight (based on experience, divided into three levels: 1, 2, and 3, the smaller the number, the higher the prosperity); the prosperity of the arrival airport of the passenger's flight (based on experience, divided into three levels: 1, 2, and 3, the smaller the number, the higher the prosperity); whether the passenger's flight is during the Spring Festival travel period; whether the passenger's flight is during holidays; the number of days in advance of the departure time from the booking time of the passenger's flight; the number of days in advance of the departure time from the ticket issuance time of the passenger's flight; the ratio of the passenger's flight fare to the average fare of the same flight on the same day; the The passenger's flight discount for this trip; the cabin class of this trip; whether the passenger voluntarily upgrades the cabin class of this trip; whether the passenger voluntarily downgrades the cabin class of this trip; the paid upgrade amount for this trip; whether the passenger is traveling in a group; the flight time of this trip; the flight distance of this trip; the historical flight delay rate of this trip; whether the passenger is a key customer or a frequent flyer; the passenger's historical number of trips (within six months); the historical average upgrade rate of this passenger's trips (if the passenger has traveled multiple times within six months); the historical average airfare discount of this passenger's trips (if the passenger has traveled multiple times within six months); the historical average paid upgrade amount of this passenger's trips (if the passenger has traveled multiple times within six months); the historical average flight delay rate of this passenger's trips (if the passenger has traveled multiple times within six months).

[0036] Step S102, determining the passenger type of the target passenger according to the passenger information, and selecting a probability prediction model according to the passenger type to obtain a target probability prediction model, wherein the target probability prediction model is used to predict the probability of the target passenger applying for an upgrade service.

[0037] It should be noted that since the data between different passengers vary greatly, in order to improve the prediction accuracy, different passengers can be classified according to their information, and the model can be trained using the data of the same type of passengers to obtain prediction models for different passenger types.

[0038] Specifically, after obtaining the passenger data, since different types of passengers require different models for prediction operations, it is necessary to determine the passenger type based on the passenger information of the target passenger, and determine the target probability prediction model among multiple prediction models based on the passenger type, and then use the target probability prediction model to process the passenger data of the target passenger to obtain the prediction result.

[0039] Step S103, inputting passenger data into the target probability prediction model, obtaining the prediction result of the probability of the target passenger applying for the upgrade service, and determining the processing result of the promotion information according to the prediction result, wherein the promotion information is used to promote the upgrade service to the target passenger.

[0040] Specifically, after obtaining the target probability prediction model and passenger data, the passenger data can be input into the target probability prediction model, and the target probability prediction model generates a predicted probability value of the target passenger's upgrade based on the passenger data, and then determines whether to send promotional information related to the upgrade to the passenger based on the predicted probability value.

[0041] For example, when passenger data is input into the target probability prediction model and the prediction result is 80%, it means that the target passenger has an 80% probability of choosing to upgrade. At this time, due to the large prediction probability value, the promotion information of the upgrade service is promoted to the passenger, thereby completing the precise promotion operation of the upgrade service and achieving the technical effect of improving the promotion efficiency of the upgrade service.

[0042] The method for processing promotion information provided by the embodiment of the present application obtains the passenger information of the target passenger and processes the passenger information numerically to obtain passenger data; determines the passenger type of the target passenger according to the passenger information, and selects a probability prediction model according to the passenger type to obtain a target probability prediction model, wherein the target probability prediction model is used to predict the probability of the target passenger applying for an upgrade service; inputs the passenger data into the target probability prediction model to obtain a prediction result of the probability of the target passenger applying for an upgrade service, and determines the processing result of the promotion information according to the prediction result, wherein the promotion information is used to promote the upgrade service to the target passenger. The problem of low promotion efficiency of the upgrade service in the related technology is solved. By processing the passenger information and determining the passenger type according to the passenger information, a model for predicting the probability of the upgrade of the passenger is determined, and the target passenger's upgrade probability is predicted by the target probability prediction model, and then it is determined whether to promote the upgrade service to the user according to the upgrade probability, thereby achieving the technical effect of improving the promotion efficiency of the upgrade service.

[0043] In order to ensure that the passenger information can be accurately identified by the model, optionally, Figure 2 is a flow chart of performing numerical processing of data according to an embodiment of the present application, such as Figure 2 As shown, in the method for processing promotion information provided in the embodiment of the present application, the passenger information includes M characteristic values, the passenger information of the target passenger is obtained, and the passenger information is numerically processed to obtain passenger data including:

[0044] Step S201, obtaining non-numerical feature values ​​in passenger information to obtain N first feature values, where M and N are positive integers, and M is equal to N.

[0045] It should be noted that since the passenger information includes both numerical feature values ​​that can be directly used for calculation (such as age, number of trips, etc.) and non-numerical feature values ​​that need to be converted before calculation (such as gender, cabin class, ticket purchase channel, etc.), in order to facilitate the subsequent target probability prediction model to process the passenger information, it is necessary to first convert the non-numerical feature values ​​in the passenger information into numerical form, thereby ensuring that all passenger information can be stored and processed in a unified numerical format.

[0046] Specifically, non-numerical feature values ​​can be feature values ​​that cannot be directly used for mathematical calculations, such as text, categories, etc., and numerical feature values ​​can be feature values ​​that can be directly used for mathematical calculations, such as integers, floating-point numbers, etc. The first feature value refers to the non-numerical feature value extracted from the passenger information, and the updated feature value refers to the first feature value in numerical form obtained after numerical processing. After obtaining the passenger information, the passenger information can be screened first to obtain the feature value that needs to be numerated from the passenger information, that is, the non-numerical feature value (first feature value).

[0047] Step S202, obtain the numerical type of each first eigenvalue, and process each first eigenvalue according to the processing method corresponding to the numerical type to obtain N updated eigenvalues, where each updated eigenvalue is a numerical value; combine the N updated eigenvalues ​​and the numerical eigenvalue in the passenger information to obtain passenger data.

[0048] Specifically, after obtaining a plurality of first eigenvalues, it is necessary to obtain the numerical type of each first eigenvalue, so as to determine the digitization method according to the numerical type, thereby ensuring that each non-numerical eigenvalue can be accurately digitized.

[0049] Furthermore, after determining the numerical processing method for each non-numerical feature value, each processing method can be used to digitize the corresponding non-numerical feature value to obtain N updated feature values. Then, the unprocessed data in the passenger information and the N updated feature values ​​can be combined to obtain the digitized passenger information, that is, the complete passenger data. At this time, all features are expressed in numerical form, which is convenient for subsequent data analysis and model training.

[0050] It should be noted that the processing methods that can be used to convert non-numerical feature values ​​into numerical values ​​include: Label Encoding: Applicable to ordered or unordered category data, each category is mapped to a unique integer. One-Hot Encoding: Convert each category into a binary vector with only one bit as 1 (indicating the category) and the rest as 0. Continuous value mapping (such as mapping to a specific interval): For some features that can be converted to continuous values ​​according to certain rules, corresponding conversion is performed.

[0051] This embodiment converts non-numerical feature values ​​into numerical values, thereby ensuring that the subsequent model can accurately identify passenger characteristics, thereby improving the prediction accuracy of the model.

[0052] Optionally, in the method for processing promotion information provided in the embodiment of the present application, each first feature value is processed according to the processing method corresponding to the numerical type to obtain N updated feature values, including: when the numerical type of the first feature value is the first type, obtaining a comparison table of the first feature value, and determining the level of the feature value according to the comparison table, wherein the first type represents that the feature value has a corresponding level, and the comparison table contains multiple feature value intervals and the level of each feature value interval; using the level to replace the first feature value to obtain an updated feature value; when the numerical type of the first feature value is the second type, encoding the first feature value by a preset encoding method to obtain an updated feature value, wherein the second type represents that the feature value does not have a corresponding level.

[0053] It should be noted that, when the eigenvalue is of the first type, the eigenvalue characterizing the type has a corresponding level, for example, when the eigenvalue is 50%, the characterization level is 2, and when the eigenvalue is 90%, the characterization level is 3. When the eigenvalue is of the second type, the eigenvalue characterizing the type does not have a corresponding level and can only be represented by text. Since different types of eigenvalues ​​can be represented in different ways, different processing methods are required to digitize non-numerical eigenvalues ​​to ensure that each non-numerical eigenvalue can obtain an accurate digitized eigenvalue.

[0054] Specifically, when determining the processing method corresponding to each numerical type, when the eigenvalue is of the first type, since the eigenvalue has a corresponding interval, each interval corresponds to a different level. Therefore, the eigenvalue interval in which the eigenvalue is located can be determined according to the comparison table of each eigenvalue, and the non-numerical eigenvalue can be digitized according to the level corresponding to the eigenvalue interval.

[0055] For example, Table 1 is an optional comparison table that can characterize the prosperity of a city. When the city in the passenger information is city A, the corresponding numerical feature value can be determined to be 1, thereby accurately determining the numerical data of the feature value.

[0056] Table 1

[0057]

[0058]

[0059] Furthermore, when the eigenvalue is of the second type, since the eigenvalue has no corresponding level, the non-numerical eigenvalue can be encoded using label encoding, one-hot encoding, continuous value mapping encoding, etc. to obtain a numerical eigenvalue.

[0060] This embodiment determines the processing method for digitizing the feature value according to the numerical type of the feature value, thereby improving the efficiency of the digitization processing and ensuring the accuracy of the digitization operation.

[0061] In order to improve the accuracy of the selected model, optionally, a probability prediction model is selected according to the passenger type to obtain a target probability prediction model, including: obtaining the frequency of target passengers taking flights in a historical time period in a database to obtain a target frequency; determining the frequency interval to which the target frequency belongs, and determining the probability prediction model associated with the frequency interval as the target probability prediction model, wherein each probability prediction model is associated with a different frequency interval.

[0062] Specifically, when determining the prediction model corresponding to the target passenger, the model can be determined according to the frequency of the passenger's flight. When constructing the model, the passenger travel frequency can be divided into three types of scales: high frequency, medium frequency, and low frequency according to the passenger travel situation of each airline in the past year. Passenger data of the same scale can be used to jointly establish a prediction model. Therefore, when determining the model used by the passenger, the frequency of the target passenger's flight in the historical time period can be obtained from the database first. When calculating the frequency, the number of times the target passenger took a flight in the historical time period can be obtained, and the frequency can be calculated based on the number of times, thereby obtaining the target frequency.

[0063] Furthermore, after obtaining the target frequency, the frequency interval in which the target frequency is located can be determined, and the probability prediction model corresponding to the frequency interval can be obtained to obtain the target probability prediction model, thereby ensuring the matching between the target passenger and the probability prediction model.

[0064] For example, Table 2 may be an optional comparison table of frequencies and models. As shown in Table 2, when the target frequency of the target passenger is 80%, the frequency may be determined to be a high frequency according to Table 2.

[0065] Table 2

[0066] frequency Frequency level 0-50％ Low Frequency 50％-80％ IF 80％-100％ high frequency

[0067] This embodiment determines the corresponding probability prediction model by determining the frequency of passengers taking flights, thereby improving the accuracy of the selected model and further improving the accuracy of the predicted upgrade probability.

[0068] In order to accurately train the model, optionally, in the method for processing promotion information provided in the embodiment of the present application, the target probability prediction model is trained in the following manner: historical passengers of the same passenger type as the target passenger in each institution are obtained to obtain P historical passenger sets, wherein each historical passenger set contains Q historical passengers, and P and Q are positive integers; historical passenger information and upgrade labels of each historical passenger in the P historical passenger sets are obtained to obtain P sample data sets, wherein each sample data set contains multiple sample data, each sample data contains historical passenger information and upgrade labels reserved for a historical passenger when taking a flight, and the upgrade label is determined according to whether the historical passenger selects the upgrade service each time he takes the flight; the initial prediction model is trained respectively by each sample data set to obtain P candidate prediction models; the model parameters in the initial prediction model are updated by the model parameters in the P candidate prediction models to obtain a first prediction model, and the training set of each candidate prediction model is updated, and the model parameters in the first prediction model are iteratively calculated by updating each candidate prediction model until the first prediction model converges to obtain the target probability prediction model.

[0069] Specifically, the model used in this application can be a federated model. Since the data between different airlines need to be kept confidential, when obtaining sample data, it is necessary to obtain it by institution, and each institution corresponds to an airline. In the case of P institutions, it is necessary to obtain P historical passenger sets, where each historical passenger set contains Q historical passengers in the institution, and obtain at least one flight data of each historical passenger, where each flight data is the user data and flight data of a passenger when he takes a flight once, as well as an upgrade label.

[0070] It should be noted that the upgrade label can be determined based on whether each passenger is upgraded each time they take a flight. For example, if the passenger is upgraded during this trip, the label is 1; if the passenger is downgraded during this trip, the label is 0; if the passenger is neither upgraded nor downgraded during this trip, the label is 0.33.

[0071] Furthermore, after determining P historical passenger sets, each historical passenger set can be used to train the initial prediction model to obtain P candidate prediction models, and the data in the P candidate prediction models can be used to update the parameters in the initial prediction model to obtain the first prediction model, and the P historical passenger sets are updated to repeat the model training to obtain new model parameters, and the first prediction model is updated using M new model parameters, and the above iterative process is repeated. When the first prediction model converges, the first prediction model can be determined as the target probability prediction model.

[0072] For example, there are 5 airlines in total, and the historical passenger set of each airline contains 100 passengers. Each passenger has data of 1-5 flights. The data of each passenger taking a flight each time is taken as a sample data, where the sample data includes passenger information, flight information and upgrade label.

[0073] It should be noted that when training the model, it is first necessary to select a suitable machine learning or deep learning model as the basic model of federated learning based on the needs and data characteristics of upgrade prediction, such as neural network, random forest, logistic regression, etc. Furthermore, a global model can be initialized on the central server, including the initial parameters and architecture of the model, and the basic parameters of federated learning can be set, such as the number of passenger data participating in the training, the number of iterations per round of training, and the size of the local data set of each institution.

[0074] Furthermore, the global model and its parameters are distributed to the local servers or devices of each airline, and the global model is trained locally using the flight data of passengers in each airline. During the local training process, each airline calculates the predicted value through forward propagation based on the global model parameters and local data, calculates the gradient through back propagation, and updates the parameters of the local model. After completing the local training, each airline uploads the parameters of the local model (such as weights and biases) to the central server. The central server collects the model parameters of each airline, updates the parameters of the global model through an aggregation algorithm (such as weighted average), and repeats the above process of distributing the global model, local training, and parameter aggregation until the predefined stopping conditions are met (such as reaching a certain number of iterations, the performance of the global model on the validation set reaches a preset threshold, etc.). The trained federated model is further evaluated using an independent test set to verify its accuracy and generalization ability in the upgrade prediction task, thereby ensuring the accuracy of the model.

[0075] Optionally, in the method for processing promotion information provided in an embodiment of the present application, historical passenger information and upgrade tags of each historical passenger in P historical passenger sets are obtained to obtain P sample data sets, including: determining the historical passenger information and upgrade tags of each historical passenger each time he takes a flight as initial sample data to obtain Q initial sample data; judging whether there are missing values ​​in each initial sample data, and if there are missing values ​​in the target initial sample data, obtaining the number of missing values ​​to obtain a target number; if the target number is greater than a preset number, deleting the target initial sample data; if the target number is less than or equal to the preset number, filling the missing values ​​to obtain updated target initial sample data; after processing each of the Q initial sample data, taking the Q updated initial sample data as a sample data set to obtain P sample data sets.

[0076] Specifically, when obtaining the historical passenger information and upgrade labels of each historical passenger each time they took a flight, some data may be missing. At this time, it is necessary to determine the number of missing data in each piece of data. If the number is too large (greater than the preset number), it means that the data cannot be used and needs to be deleted. If the number is less than or equal to the preset number, the missing values ​​can be filled to obtain the updated target initial sample data, thereby ensuring that the number of sample data meets the training requirements while ensuring the quality of the sample data.

[0077] Optionally, in the method for processing promotion information provided in an embodiment of the present application, filling missing values ​​to obtain updated target initial sample data includes: when the missing value is categorical, replacing the missing value with a preset value; when the missing value is numerical, obtaining the target feature value corresponding to the missing value in the initial sample data with the highest similarity to the target initial sample data, and replacing the missing value with the target feature value.

[0078] Specifically, when the missing type is categorical, since it is unknown what the category is, a preset value, such as -1, can be used directly to fill in the position of the missing value. When the missing value is numerical, the missing value can be filled with the value of the missing position in the initial sample data with the highest similarity to the sample data, thereby ensuring the accuracy of filling the missing value.

[0079] For example, to handle missing values ​​and outliers, the following two methods are used according to the specific situation of each missing value:

[0080] (1) For categorical missing values, the default value -1 is filled, indicating that the missing value is missing or abnormal;

[0081] (2) For missing values ​​of numerical type, find a sample similar to it in the remaining sample data, and then fill the missing value with the corresponding characteristic value of the sample.

[0082] This embodiment achieves the technical effect of improving the accuracy of sample data by filling in missing values ​​in sample data.

[0083] Optionally, in the method for processing promotion information provided in an embodiment of the present application, determining the processing result of the promotion information based on the prediction result includes: when the prediction result indicates that the probability of the target passenger being upgraded is greater than or equal to a preset probability, executing the operation of sending the promotion information to the target passenger; when the prediction result indicates that the probability of the target passenger being upgraded is less than a preset probability, not executing the operation of sending the promotion information to the target passenger.

[0084] Specifically, when the prediction result indicates that the target passenger's upgrade probability is greater than or equal to the preset probability, it indicates that the passenger has a high willingness to upgrade. At this time, upgrade promotion information can be pushed to the passenger, so that the passenger can obtain upgrade-related business information in a timely manner and perform upgrade operations. When the prediction result indicates that the target passenger's upgrade probability is less than the preset probability, it indicates that the passenger has a low willingness to upgrade, and there is no need to push upgrade promotion information to the passenger, thereby avoiding affecting the passenger's normal itinerary and affecting the passenger's flight experience.

[0085] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0086] The embodiment of the present application also provides a promotion information processing device. It should be noted that the promotion information processing device of the embodiment of the present application can be used to execute the promotion information processing method provided in the embodiment of the present application. The promotion information processing device provided in the embodiment of the present application is introduced below.

[0087] Figure 3 Schematic diagram of a device for processing promotion information provided according to an embodiment of the present application. Figure 3 As shown, the device includes: a first acquisition unit 31, a determination unit 32, and a prediction unit 33.

[0088] The first acquisition unit 31 is used to acquire the passenger information of the target passenger and perform numerical processing on the passenger information to obtain passenger data.

[0089] The determination unit 32 is used to determine the passenger type of the target passenger according to the passenger information, and select a probability prediction model according to the passenger type to obtain a target probability prediction model, wherein the target probability prediction model is used to predict the probability of the target passenger applying for an upgrade service.

[0090] The prediction unit 33 is used to input passenger data into the target probability prediction model to obtain the prediction result of the probability of the target passenger applying for the upgrade service, and determine the processing result of the promotion information according to the prediction result, wherein the promotion information is used to promote the upgrade service to the target passenger.

[0091] The promotion information processing device provided in the embodiment of the present application obtains the passenger information of the target passenger through the first acquisition unit 31, and digitizes the passenger information to obtain passenger data; the determination unit 32 determines the passenger type of the target passenger according to the passenger information, and selects a probability prediction model according to the passenger type to obtain a target probability prediction model, wherein the target probability prediction model is used to predict the probability of the target passenger applying for an upgrade service; the prediction unit 33 inputs the passenger data into the target probability prediction model to obtain a prediction result of the probability of the target passenger applying for an upgrade service, and determines the processing result of the promotion information according to the prediction result, wherein the promotion information is used to promote the upgrade service to the target passenger, thereby solving the problem of low promotion efficiency of the upgrade service in the related art, by processing the passenger information and determining the passenger type according to the passenger information, thereby determining a model for predicting the upgrade probability of the passenger, and predicting the upgrade probability of the target passenger through the target probability prediction model, and then determining whether to promote the upgrade service to the user according to the upgrade probability, thereby achieving the technical effect of improving the promotion efficiency of the upgrade service.

[0092] Optionally, in the promotion information processing device provided in the embodiment of the present application, the passenger information includes M characteristic values, and the first acquisition unit 31 includes: a first acquisition module, used to obtain non-numerical characteristic values ​​in the passenger information to obtain N first characteristic values, wherein M and N are positive integers, and M is equal to N; a second acquisition module, used to obtain the numerical type of each first characteristic value, and process each first characteristic value according to the processing method corresponding to the numerical type to obtain N updated characteristic values, wherein each updated characteristic value is a numerical value; a combination module, used to combine the N updated characteristic values ​​with the numerical characteristic values ​​in the passenger information to obtain passenger data.

[0093] Optionally, in the promotion information processing device provided in the embodiment of the present application, the second acquisition module includes: a determination submodule, which is used to obtain a comparison table of the first feature value when the numerical type of the first feature value is the first type, and determine the level of the feature value according to the comparison table, wherein the first type represents that the feature value has a corresponding level, and the comparison table contains multiple feature value intervals and the level of each feature value interval; a first replacement submodule, which is used to replace the first feature value with the level to obtain an updated feature value; an encoding submodule, which is used to encode the first feature value by a preset encoding method to obtain an updated feature value when the numerical type of the first feature value is the second type, wherein the second type represents that the feature value does not have a corresponding level.

[0094] Optionally, in the promotion information processing device provided in the embodiment of the present application, the determination unit 32 includes: a third acquisition module, used to obtain the frequency of target passengers taking flights in a historical time period in the database to obtain a target frequency; a first determination module, used to determine the frequency interval to which the target frequency belongs, and determine the probability prediction model associated with the frequency interval as the target probability prediction model, wherein each probability prediction model is associated with a different frequency interval.

[0095] Optionally, in the promotion information processing device provided in the embodiment of the present application, the target probability prediction model is trained and obtained in the following manner: a second acquisition unit is used to acquire historical passengers of the same passenger type as the target passenger in each institution, and obtain P historical passenger sets, wherein each historical passenger set contains Q historical passengers, and P and Q are positive integers; a third acquisition unit is used to acquire historical passenger information and upgrade labels of each historical passenger in the P historical passenger sets, and obtain P sample data sets, wherein each sample data set contains multiple sample data, and each sample data contains historical passenger information and upgrade labels reserved for a historical passenger when taking a flight, and the upgrade label is determined according to whether the historical passenger selects the upgrade service each time he takes the flight; a training unit is used to train the initial prediction model respectively through each sample data set to obtain P candidate prediction models; an iteration unit is used to update the model parameters in the initial prediction model through the model parameters in the P candidate prediction models to obtain a first prediction model, and update the training set of each candidate prediction model, and iteratively calculate the model parameters in the first prediction model by updating each candidate prediction model until the first prediction model converges to obtain the target probability prediction model.

[0096] Optionally, in the promotion information processing device provided in the embodiment of the present application, the third acquisition unit includes: a second determination module, used to determine the historical passenger information and upgrade label of each historical passenger each time he takes a flight as initial sample data, and obtain Q initial sample data; a judgment module, used to judge whether there are missing values ​​in each initial sample data, and when there are missing values ​​in the target initial sample data, obtain the number of missing values ​​to obtain the target number; a deletion module, used to delete the target initial sample data when the target number is greater than a preset number; a filling module, used to fill the missing values ​​when the target number is less than or equal to the preset number to obtain updated target initial sample data; a processing module, used to take the Q updated initial sample data as a sample data set after processing each of the Q initial sample data, and obtain P sample data sets.

[0097] Optionally, in the processing device for promotion information provided in an embodiment of the present application, the filling module includes: a second replacement submodule, used to replace the missing value with a preset value when the missing value is a categorical type; a third replacement submodule, used to obtain the target feature value corresponding to the missing value in the initial sample data with the highest similarity to the target initial sample data when the missing value is a numerical type, and replace the missing value with the target feature value.

[0098] Optionally, in the promotion information processing device provided in the embodiment of the present application, the prediction unit 33 includes: a first execution module, used to execute the operation of sending promotion information to the target passenger when the prediction result indicates that the probability of the target passenger's upgrade is greater than or equal to a preset probability; and a second execution module, used to not execute the operation of sending the promotion information to the target passenger when the prediction result indicates that the probability of the target passenger's upgrade is less than a preset probability.

[0099] The above-mentioned promotion information processing device includes a processor and a memory. The above-mentioned first acquisition unit 31, determination unit 32, prediction unit 33, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions.

[0100] The processor includes a kernel, and the kernel calls the corresponding program unit from the memory. One or more kernels can be set, and the problem of low promotion efficiency of the upgrade business in the related art is solved by adjusting the kernel parameters.

[0101] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0102] An embodiment of the present invention provides a computer-readable storage medium on which a program is stored. When the program is executed by a processor, the promotion information processing method is implemented.

[0103] An embodiment of the present invention provides a processor, which is used to run a program, wherein the promotion information processing method is executed when the program is running.

[0104] Figure 4 is a schematic diagram of an electronic device provided according to an embodiment of the present application, such as Figure 4 As shown, an embodiment of the present invention provides an electronic device, the electronic device 40 includes a processor, a memory, and a program stored in the memory and executable on the processor, and the processor implements the steps of the above-mentioned promotion information processing method when executing the program. The device in this article can be a server, a PC, a PAD, a mobile phone, etc.

[0105] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that initializes the steps of the above-mentioned promotion information processing method.

[0106] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0107] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0108] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0109] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0110] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0111] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0112] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0113] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0114] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for processing promotion information, characterized in that: include: Acquire the passenger information of the target passenger, and digitize the passenger information to obtain passenger data; Determining the passenger type of the target passenger according to the passenger information, and selecting a probability prediction model according to the passenger type to obtain a target probability prediction model, wherein the target probability prediction model is used to predict the probability of the target passenger applying for an upgrade service; The passenger data is input into the target probability prediction model to obtain a prediction result of the probability of the target passenger applying for the upgrade service, and a processing result of the promotion information is determined according to the prediction result, wherein the promotion information is used to promote the upgrade service to the target passenger.

2. The method according to claim 1, characterized in that The passenger information includes M characteristic values, and the passenger information of the target passenger is obtained, and the passenger information is numerically processed to obtain passenger data including: Obtaining non-numerical feature values ​​in the passenger information to obtain N first feature values, where M and N are positive integers, and M is equal to N; Obtaining the numerical type of each first eigenvalue, and processing each first eigenvalue according to a processing method corresponding to the numerical type to obtain N updated eigenvalues, wherein each updated eigenvalue is a numerical value; The N updated feature values ​​are combined with the numerical feature value in the passenger information to obtain the passenger data.

3. The method according to claim 2, characterized in that Each first eigenvalue is processed according to the processing method corresponding to the numerical type to obtain N updated eigenvalues ​​including: In the case where the numerical type of the first eigenvalue is a first type, obtaining a comparison table of the first eigenvalue, and determining the level of the eigenvalue according to the comparison table, wherein the first type represents that the eigenvalue has a corresponding level, and the comparison table includes multiple eigenvalue intervals and the level of each eigenvalue interval; Using the level to replace the first feature value, to obtain the updated feature value; In the case where the numerical type of the first eigenvalue is the second type, the first eigenvalue is encoded by a preset encoding method to obtain the updated eigenvalue, wherein the second type indicates that the eigenvalue does not have a corresponding level.

4. The method according to claim 1, characterized in that: Selecting a probability prediction model according to the passenger type to obtain a target probability prediction model includes: Obtaining the frequency of the target passenger taking an airplane in a historical period in a database to obtain a target frequency; The frequency interval to which the target frequency belongs is determined, and a probability prediction model associated with the frequency interval is determined as the target probability prediction model, wherein each probability prediction model is associated with a different frequency interval.

5. The method according to claim 1, characterized in that The target probability prediction model is trained in the following way: Obtain historical passengers of the same passenger type as the target passenger in each institution to obtain P historical passenger sets, wherein each historical passenger set contains Q historical passengers, and P and Q are positive integers; Obtaining historical passenger information and an upgrade tag of each historical passenger in the P historical passenger sets, to obtain P sample data sets, wherein each sample data set includes a plurality of sample data, each sample data includes historical passenger information and an upgrade tag reserved when a historical passenger takes a flight, and the upgrade tag is determined according to whether the historical passenger selects an upgrade service each time when taking a flight; The initial prediction model is trained through each sample data set to obtain P candidate prediction models; The model parameters in the initial prediction model are updated by the model parameters in the P candidate prediction models to obtain a first prediction model, and the training set of each candidate prediction model is updated. The model parameters in the first prediction model are iteratively calculated by updating each candidate prediction model until the first prediction model converges to obtain the target probability prediction model.

6. The method according to claim 5, characterized in that The historical passenger information and upgrade label of each historical passenger in the P historical passenger sets are obtained to obtain P sample data sets including: The historical passenger information and upgrade label of each historical passenger each time he / she takes an airplane are determined as initial sample data, and Q initial sample data are obtained; Determine whether there are missing values ​​in each initial sample data, and if there are missing values ​​in the target initial sample data, obtain the number of missing values ​​to obtain the target number; When the target quantity is greater than a preset quantity, deleting the target initial sample data; When the target number is less than or equal to the preset number, fill in the missing values ​​to obtain updated target initial sample data; After processing each of the Q initial sample data, the Q updated initial sample data are taken as a sample data set to obtain the P sample data sets.

7. The method according to claim 6, characterized in that Filling the missing values ​​to obtain updated target initial sample data includes: When the missing value is categorical, the missing value is replaced by a preset value; In the case where the missing value is of numerical type, a target feature value corresponding to the missing value in the initial sample data having the highest similarity to the target initial sample data is obtained, and the missing value is replaced by the target feature value.

8. The method according to claim 1, characterized in that Determining the processing result of the promotion information according to the prediction result includes: When the prediction result indicates that the probability of the target passenger being upgraded is greater than or equal to a preset probability, sending the promotion information to the target passenger; When the prediction result indicates that the probability of the target passenger being upgraded is less than the preset probability, the operation of sending the promotion information to the target passenger is not performed.

9. A promotion information processing device, characterized in that: include: A first acquisition unit is used to acquire passenger information of a target passenger and digitize the passenger information to obtain passenger data; a determination unit, configured to determine the passenger type of the target passenger according to the passenger information, and select a probability prediction model according to the passenger type to obtain a target probability prediction model, wherein the target probability prediction model is used to predict the probability of the target passenger applying for an upgrade service; A prediction unit is used to input the passenger data into the target probability prediction model to obtain a prediction result of the probability of the target passenger applying for the upgrade service, and determine a processing result of the promotion information according to the prediction result, wherein the promotion information is used to promote the upgrade service to the target passenger.

10. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the promotion information processing method described in any one of claims 1 to 8.

Images