Airline ticket transaction prediction method and apparatus

Abstract

The present disclosure provides an airline ticket transaction prediction method and apparatus. The method comprises: acquiring an airline ticket transaction log, and extracting a plurality of transaction data groups from the transaction log; processing each transaction data group, and distinguishing first-type transaction data and second-type transaction data in each transaction data group; training an airline ticket prediction model on the basis of each piece of first-type transaction data and each piece of second-type transaction data; and when an airline ticket transaction prediction request corresponding to a target flight is received, acquiring transaction data to be predicted corresponding to the airline ticket transaction prediction request, and inputting the transaction data to be predicted into the airline ticket prediction model to obtain a prediction result output by the airline ticket prediction model, wherein the prediction result is used for feeding back the willingness degree of purchasing the target flight by a customer. By applying the method provided in the present disclosure, the willingness of purchasing an airline ticket by a customer can be predicted on the basis of a plurality of pieces of transaction data, thereby facilitating subsequent formulation of related ticket selling policies.

Description

Air ticket transaction prediction method and device

[0001] This disclosure claims priority to the Chinese patent application filed with the China Patent Office on November 30, 2023, with application number 202311636320.1 and application name “Air Ticket Transaction Prediction Method and Device,” the entire contents of which are incorporated by reference into this disclosure. Technical Field

[0002] The present disclosure relates to the field of computer technology, and in particular to a method and device for predicting air ticket transactions. Background Art

[0003] In recent years, the tourism market has experienced rapid growth, with online booking and ticket purchase becoming the preferred method for many travelers. Passengers have numerous online ticket purchasing channels, including major airline websites and online travel agencies (OTAs). With increasing price transparency, various airlines are competing to gain market share in the civil aviation market, launching distinctive airfare products. This has resulted in a wide variety of airfares and services to meet the needs of diverse travelers.

[0004] Merchants offer flexible and diverse customized air ticket products, primarily through price and service differentiation. Ultimately, whether a transaction is completed depends on the passenger's willingness to purchase the product. This willingness is influenced by numerous factors, including price, service, time, inventory, and competing products. Among these factors, airfare plays a decisive role. It can be said that airfare and passenger purchase willingness are closely related and can be considered equivalent. When airfares rise, passenger purchase willingness decreases; when airfares fall, passenger purchase willingness increases. Therefore, during air ticket transactions, merchants need to balance these two factors, customizing airfares that meet passenger purchase preferences. However, currently, it is difficult for merchants to strike a balance between these two factors, making it impossible to determine passenger purchase willingness under given conditions. Summary of the Invention

[0005] In view of this, the present disclosure provides an air ticket transaction prediction method, through which the customer's willingness to purchase air tickets can be predicted based on multiple transaction data, and it can be determined whether all tickets for the target flight have been successfully sold, so as to facilitate the subsequent formulation of relevant ticket sales policies.

[0006] The present disclosure also provides an air ticket transaction prediction device to ensure the implementation and application of the above method in practice.

[0007] A method for predicting air ticket transactions, comprising:

[0008] Obtaining an air ticket transaction log, and capturing multiple transaction data groups in the transaction log;

[0009] processing each of the transaction data groups to distinguish between first-category transaction data and second-category transaction data in each of the transaction data groups;

[0010] Training an air ticket prediction model based on each of the first-category transaction data and the second-category transaction data;

[0011] When a ticket transaction prediction request corresponding to a target flight is received, the transaction data to be predicted corresponding to the ticket transaction prediction request is obtained, and the transaction data to be predicted is input into the ticket prediction model to obtain the prediction result output by the ticket prediction model. The prediction result is used to provide feedback on whether all tickets for the target flight have been successfully sold.

[0012] In the above method, optionally, the capturing of multiple transaction data groups in the transaction log includes:

[0013] Get the transaction serial number within the preset historical time period;

[0014] Searching the transaction log for an initial transaction data group corresponding to each transaction serial number, wherein the initial transaction data group includes a plurality of initial transaction data;

[0015] Data cleaning is performed on each of the initial transaction data groups to obtain a plurality of cleaned transaction data groups.

[0016] Optionally, the method described above processes each of the transaction data groups to distinguish between the first type of transaction data and the second type of transaction data in each of the transaction data, including:

[0017] Based on a preset data table, determining transaction data belonging to direct influencing factors and transaction data belonging to indirect influencing factors in each of the transaction data groups;

[0018] Acquire data parameters corresponding to the transaction data of each of the directly influencing factors as characteristic values ​​of the transaction data, and determine that the transaction data of the directly influencing factors are first-category transaction data;

[0019] Data parameters corresponding to the transaction data belonging to the indirect influencing factors are split and reorganized to obtain characteristic values ​​of each transaction data belonging to the indirect influencing factors, and the transaction data belonging to the indirect influencing factors are determined to be second-category transaction data.

[0020] Optionally, in the above method, the step of training the air ticket prediction model based on each of the first-category transaction data and the second-category transaction data includes:

[0021] Obtaining a preset training target corresponding to each of the transaction data groups, determining a first data weight for each piece of the first-category transaction data and a second data weight for each piece of the second-category transaction data in each of the transaction data groups, and generating an actual data weight corresponding to each of the transaction data groups based on the first data weights and the second data weights;

[0022] Using the characteristic value of each of the first-category transaction data and the characteristic value of the second-category transaction data as training data, and using the weight of each of the actual data as a model parameter of the air ticket prediction model;

[0023] Inputting each training data group into the air ticket prediction model, triggering the air ticket prediction model to perform iterative training based on the SVM algorithm and each training data;

[0024] Applying the training target corresponding to each transaction data group to verify the training results of each training of the air ticket prediction model;

[0025] If the current training result meets the training goal, the training of the air ticket prediction model is completed.

[0026] Optionally, the method described above includes inputting the transaction data to be predicted into the air ticket prediction model to obtain a prediction result output by the air ticket prediction model, including:

[0027] Obtaining characteristic values ​​of the transaction data to be predicted;

[0028] The characteristic values ​​of the transaction data to be predicted are input into the air ticket prediction model to obtain the prediction result output by the air ticket prediction model.

[0029] An air ticket transaction prediction device, comprising:

[0030] An acquiring unit, configured to acquire a ticket transaction log and capture a plurality of transaction data groups in the transaction log;

[0031] a processing unit configured to process each of the transaction data groups and distinguish between first-category transaction data and second-category transaction data in each of the transaction data groups;

[0032] a training unit configured to train an air ticket prediction model based on each of the first-category transaction data and the second-category transaction data;

[0033] The prediction unit is configured to, when receiving a ticket transaction prediction request corresponding to a target flight, obtain the transaction data to be predicted corresponding to the ticket transaction prediction request, input the transaction data to be predicted into the ticket prediction model, and obtain the prediction result output by the ticket prediction model, wherein the prediction result is used to provide feedback on whether all tickets for the target flight are successfully sold.

[0034] In the above device, optionally, the acquisition unit includes:

[0035] The first acquisition subunit is configured to acquire transaction serial numbers within a preset historical time period;

[0036] a search subunit configured to search the transaction log for an initial transaction data group corresponding to each transaction serial number, wherein the initial transaction data group includes a plurality of initial transaction data;

[0037] The data cleaning subunit is configured to clean each initial transaction data in each of the initial transaction data groups to obtain a plurality of cleaned transaction data groups.

[0038] In the above device, optionally, the processing unit includes:

[0039] a first determining subunit configured to determine, based on a preset data table, transaction data belonging to direct influencing factors and transaction data belonging to indirect influencing factors in each of the transaction data groups;

[0040] a second acquiring subunit configured to acquire data parameters corresponding to the transaction data of each of the directly influencing factors as characteristic values ​​of the transaction data, and determine that the transaction data of the directly influencing factors are first-category transaction data;

[0041] The processing subunit is configured to split and reorganize the data parameters corresponding to the transaction data belonging to the indirect influencing factors, obtain the characteristic value of each transaction data belonging to the indirect influencing factors, and determine that the transaction data belonging to the indirect influencing factors is the second category of transaction data.

[0042] In the above-mentioned device, optionally, the training unit includes:

[0043] a second determining subunit, configured to obtain a preset training target corresponding to each of the transaction data groups, determine a first data weight of each of the first-category transaction data and a second data weight of each of the second-category transaction data in each of the transaction data groups, and generate an actual data weight corresponding to each of the transaction data groups based on each of the first data weights and the second data weights;

[0044] a third determining subunit, configured to use the characteristic value of each of the first-category transaction data and the characteristic value of the second-category transaction data as training data, and the weight of each of the actual data as a model parameter of the air ticket prediction model;

[0045] A training subunit, which groups the training data into the air ticket prediction model, triggering the air ticket prediction model to perform iterative training based on the SVM algorithm and the training data;

[0046] A verification subunit, configured to apply the training target corresponding to each transaction data group to verify the training results of each training of the air ticket prediction model;

[0047] The training subunit is stopped and is configured to complete the training of the air ticket prediction model if the current training result meets the training target.

[0048] In the above device, optionally, the prediction unit includes:

[0049] a third acquisition subunit, configured to acquire a characteristic value of the transaction data to be predicted;

[0050] The prediction subunit is configured to input the characteristic values ​​of the transaction data to be predicted into the air ticket prediction model to obtain the prediction result output by the air ticket prediction model.

[0051] Compared with the prior art, the present disclosure has the following advantages:

[0052] The present disclosure provides a method for predicting air ticket transactions, comprising: obtaining an air ticket transaction log, capturing multiple transaction data groups in the transaction log; processing each of the transaction data groups, distinguishing between first-category transaction data and second-category transaction data in each of the transaction data groups; training an air ticket prediction model based on each of the first-category transaction data and the second-category transaction data; upon receiving an air ticket transaction prediction request corresponding to a target flight, obtaining the transaction data to be predicted corresponding to the air ticket prediction request, and inputting the transaction data to be predicted into the air ticket prediction model to obtain a prediction result output by the air ticket prediction model. By applying the method provided by the present disclosure, a customer's willingness to purchase air tickets can be predicted based on multiple transaction data, so as to facilitate the subsequent formulation of relevant ticket sales policies. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0054] FIG1 is a flow chart of a method for predicting air ticket transactions provided by an embodiment of the present disclosure;

[0055] FIG2 is another method flow chart of an air ticket transaction prediction method provided by an embodiment of the present disclosure;

[0056] FIG3 is a flow chart of another method for predicting air ticket transactions provided by an embodiment of the present disclosure;

[0057] FIG4 is a device structure diagram of an air ticket transaction prediction device provided by an embodiment of the present disclosure;

[0058] FIG5 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0059] The following will be combined with the accompanying drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Obviously, the embodiments described are only part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present disclosure.

[0060] In this application, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. The terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or apparatus comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or apparatus comprising the element.

[0061] The present disclosure can be used in a variety of general-purpose or special-purpose computing device environments or configurations, such as personal computers, server computers, handheld or portable devices, tablet devices, multi-processor devices, distributed computing environments including any of the above, and the like.

[0062] The present disclosure provides an air ticket transaction prediction method. The method can be applied to various system platforms. The execution subject can be a computer terminal or a processor of various mobile devices. The method flow chart is shown in FIG1 , which specifically includes:

[0063] S101: Acquire an air ticket transaction log, and capture multiple transaction data groups in the transaction log.

[0064] Specifically, capturing multiple transaction data groups in the transaction log includes:

[0065] Get the transaction serial number within the preset historical time period;

[0066] Searching the transaction log for an initial transaction data group corresponding to each transaction serial number, wherein the initial transaction data group includes a plurality of initial transaction data;

[0067] Data cleaning is performed on each of the initial transaction data groups to obtain a plurality of cleaned transaction data groups.

[0068] It should be noted that the ticket transaction log is from the production environment and records the transaction requests and results. The transaction request includes:

[0069] 1) Passenger information: passenger identity (adult, child, infant, etc.), number of passengers, VIP customer identity, etc.;

[0070] 2) Sales information: sales time, sales channels, sales locations, sales identities, etc.;

[0071] 3) Itinerary information: origin, destination, flight, class, etc.

[0072] Transaction results include: air ticket prices, services and other information.

[0073] The above-mentioned transaction data is captured in the transaction log for data cleaning. The request and result are matched according to the transaction serial number, and the transaction data of each transaction serial number with the same transaction serial number is generated into a transaction data group.

[0074] It should also be noted that data cleaning is to delete some unnecessary transaction data. For example, an agent requests a ticket with a departure time of 2022-10-01 / 15:00 and a ticket issuance time of 2022-09-16 / 17:15. The agent here has no impact on the transaction, while the departure time and ticket issuance time do affect the transaction. Therefore, the agent content is deleted, and the departure time and ticket issuance time content is retained.

[0075] S102: Process each of the transaction data groups to distinguish between the first type of transaction data and the second type of transaction data in each of the transaction data groups.

[0076] It should be noted that the transaction data group contains multiple transaction data. Some of the transaction data can directly affect the user's ticket purchase choice, while some transaction data indirectly affect the user's ticket purchase choice. Therefore, the transaction data is divided into the first category of transaction data and the second category of transaction data according to different influencing factors.

[0077] Specifically, the first type of transaction data is transaction data that has a direct impact on air ticket transactions; the second type of transaction data is transaction data that has an indirect impact on air tickets.

[0078] S103: Training an air ticket prediction model based on each of the first-category transaction data and the second-category transaction data.

[0079] S104: When an air ticket transaction prediction request corresponding to a target flight is received, transaction data to be predicted corresponding to the air ticket transaction prediction request is obtained, and the transaction data to be predicted is input into the air ticket prediction model to obtain a prediction result output by the air ticket prediction model.

[0080] The prediction result is used to provide feedback on whether all tickets for the target flight are successfully sold.

[0081] It should be noted that airlines need to predict the transaction status of tickets for a specific flight by sending a corresponding ticket transaction prediction request based on transaction data such as flight number, price, and flight time. After receiving this request, the processor obtains the transaction data to be predicted corresponding to the request and inputs this transaction data into the ticket prediction model. The resulting prediction result is the predicted transaction status of tickets for the target flight, that is, the predicted probability of successfully selling tickets for the target flight.

[0082] In the method provided by the embodiments of the present disclosure, transaction serial numbers within a historical time period are obtained from transaction logs. Initial historical transaction data sets are then obtained based on these transaction serial numbers. Each initial transaction data set contains individual transaction data for a particular flight at a specific point in time. Each transaction data set is processed and classified, and a flight ticket prediction model is trained based on the classified transaction data. After the model is trained, the model is applied to predict flight ticket transactions.

[0083] By applying the method provided in the embodiments of the present disclosure, a customer's willingness to purchase air tickets can be predicted based on multiple transaction data, so as to facilitate the subsequent formulation of relevant ticket sales policies.

[0084] In the method provided in the embodiment of the present disclosure, the process of processing each of the transaction data groups and distinguishing the first type of transaction data from the second type of transaction data in each of the transaction data is shown in FIG2 , and includes:

[0085] S201: Based on a preset data table, determining transaction data belonging to direct influencing factors and transaction data belonging to indirect influencing factors in each transaction data group.

[0086] It should be noted that the data table specifies which transaction data belongs to the transaction data of direct influencing factors and which data belongs to the data of indirect influencing factors.

[0087] S202: Acquire data parameters corresponding to the transaction data of each of the directly influencing factors as characteristic values ​​of the transaction data, and determine that the transaction data of the directly influencing factors are first-category transaction data.

[0088] S203: performing data splitting and reorganization on the data parameters corresponding to the transaction data belonging to the indirect influencing factors, obtaining characteristic values ​​of each transaction data belonging to the indirect influencing factors, and determining that the transaction data belonging to the indirect influencing factors are second-category transaction data.

[0089] In this disclosure, direct influencing factors refer to certain elements in transaction data that can be directly used to train models. These elements are referred to as direct influencing factors and do not require conversion. For example, the allowable luggage weight is a data parameter in the transaction data, and this parameter is used as a feature value. Indirect influencing factors refer to certain elements in the transaction data that have no direct value for training models and cannot be used directly. These elements are referred to as indirect influencing factors and require conversion. For example, the departure time and sales time mentioned above, by themselves, cannot directly generate valuable information for training. However, these factors have hidden value. If the departure time falls on a holiday, the number of tourists will increase, and airfares will increase. The longer the advance purchase period (the time difference between sales time and departure time), the lower the airfare, and the closer the departure time, the higher the airfare. Feature conversion includes operations such as splitting and reorganizing. For example, the departure time can be split into date and time, and the date can be converted into a "holiday" feature, and the time can be converted into a "flight period" feature. The sales time and departure time can be combined to form the advance purchase period. These converted features need to be populated with accurate and reasonable values. For example, the "Is it a holiday" feature only needs to be set to "1" or "0" according to the original value. "1" represents "yes" and "0" represents "no".

[0090] In the method provided by the embodiment of the present disclosure, the process of training the air ticket prediction model based on each of the first type of transaction data and the second type of transaction data is shown in FIG3 , and includes:

[0091] S301: Obtain a pre-set training target corresponding to each of the transaction data groups, and determine the first data weight of each of the first-category transaction data and the second data weight of each of the second-category transaction data in each of the transaction data groups, and generate the actual data weight corresponding to each of the transaction data groups based on each of the first data weights and the second data weights.

[0092] It should be noted that the actual data weight can be the sum of the data weights of each transaction data in the transaction data group. The first data weight is greater than the second data weight. The more first-category transaction data a transaction data group contains, the greater the actual data weight of the transaction data group.

[0093] It should also be noted that the training goal is to expect the air ticket prediction model to correctly output the expected parameters.

[0094] S302: Using the characteristic value of each of the first-category transaction data and the characteristic value of the second-category transaction data as training data, and using the weight of each of the actual data as a model parameter of the air ticket prediction model.

[0095] S303: Inputting each training data group into the air ticket prediction model, triggering the air ticket prediction model to perform iterative training based on the SVM algorithm and each training data.

[0096] It's important to note that the airfare prediction model used is a support vector machine (SVM). SVM is a binary classification model and a supervised learning method. Its basic model is a linear classifier defined by maximizing the margin in the feature space. It can also perform nonlinear classification using kernel methods. The SVM learning strategy is to maximize the margin, which can be formalized as a convex quadratic programming problem. Its learning algorithm is an optimization algorithm for solving convex quadratic programming.

[0097] S304: Apply the training target corresponding to each transaction data group to verify the training results of each training of the air ticket prediction model.

[0098] It should be noted that the training result output by the ticket prediction model each time is a prediction of whether the ticket is successfully sold.

[0099] S305: If the current training result meets the training target, the training of the air ticket prediction model is completed.

[0100] In this disclosure, we use an SVM training sample dataset. Based on the principles of the SVM algorithm, we iterate repeatedly until the maximum number of iterations or an acceptable error is reached, at which point training is stopped. Ultimately, the SVM model predicts a transaction outcome, either success or failure. A successful transaction prediction indicates that the ticket price meets the passenger's purchasing intent, while a failed transaction prediction indicates that the ticket price does not meet the passenger's purchasing intent.

[0101] It should be noted that if the current training results do not meet the training objectives, the air ticket prediction model automatically adjusts the model parameters and enters the next training.

[0102] Furthermore, inputting the transaction data to be predicted into the air ticket prediction model to obtain the prediction result output by the air ticket prediction model includes:

[0103] Obtaining characteristic values ​​of the transaction data to be predicted;

[0104] The characteristic values ​​of the transaction data to be predicted are input into the air ticket prediction model to obtain the prediction result output by the air ticket prediction model.

[0105] It should be noted that after obtaining the transaction data to be predicted, it is classified into the first category of transaction data belonging to the direct influencing factors and the second category of transaction data. The characteristic values ​​corresponding to each category of transaction data are obtained. The characteristic values ​​are input into the air ticket prediction model to obtain the output prediction results.

[0106] Based on the method provided in the above embodiment, the process of predicting the air ticket prediction model can be implemented as follows:

[0107] Through the pre-processing link, the training data for adult passengers from HYN to CKG carried by XX is obtained.

[0108] Status represents the transaction result, with "+1" representing a successful transaction and "-1" representing an unsuccessful transaction. The corresponding data for booking class, seat, equipment type, fare amount, baggage, stopover, service, is holiday, travel time range, and advance ticket transaction data are feature values. Booking class is typically defined using 26 letters, which translate to numerical values ​​from 1 to 26. "E" translates to "5." Baggage weight is standardized in kilograms throughout the industry, so the corresponding unit is removed from the data. The value "M" in travelTimeRange represents morning, so for ease of calculation, it is converted to "2." The value "33" in advance ticket purchase duration represents a ticket purchased 33 days in advance. If the data is not in days, it is converted to the corresponding numerical value; for example, 12 hours translates to 0.5 days.

[0109] In the above embodiment, each feature has two training data items (x1, y1) and (x2, y2), corresponding to the first and second data items, respectively. x1 and x2 are 10-dimensional vectors, corresponding to 10 features, and y1 = "+1" and y2 = "-1". For example, x1 and y1 correspond to the first data item in the table; x2 and y2 correspond to the second data item in the table. In combination with a practical example, x1 = {5, 8, 738, 710.00, 10, 0, 0, 0, 2, 33}, y1 = +1. This training data set is used as the input of the SVM, and w is initialized with random values, and b is initialized to 0. Iterative training begins until the two training data items are separated according to status and a prediction model is obtained. Subsequent new data is input into the prediction model, and the status is calculated. If it is "+1", it means that the air ticket price transaction is successful, and if it is "-1", it means that the air ticket price transaction failed.

[0110] The specific implementation processes and derivative methods of the above embodiments are all within the protection scope of this disclosure.

[0111] Corresponding to the method described in FIG1 , an embodiment of the present disclosure further provides an air ticket transaction prediction device, which is configured to specifically implement the method described in FIG1 . The air ticket transaction prediction device provided in the embodiment of the present disclosure can be applied to a computer terminal or various mobile devices. A schematic structural diagram thereof is shown in FIG4 , specifically comprising:

[0112] An acquisition unit 401 is configured to acquire a ticket transaction log and capture multiple transaction data groups in the transaction log;

[0113] A processing unit 402 is configured to process each of the transaction data groups and distinguish between the first type of transaction data and the second type of transaction data in each of the transaction data groups;

[0114] A training unit 403 is configured to train an air ticket prediction model based on each of the first-category transaction data and the second-category transaction data;

[0115] The prediction unit 404 is configured to, when receiving an air ticket transaction prediction request corresponding to a target flight, obtain the transaction data to be predicted corresponding to the air ticket transaction prediction request, input the transaction data to be predicted into the air ticket prediction model, and obtain the prediction result output by the air ticket prediction model, and the prediction result is used to provide feedback on the customer's willingness to purchase the target flight.

[0116] In the device provided by the embodiments of the present disclosure, transaction serial numbers within a historical time period are obtained from transaction logs, and initial historical transaction data sets are obtained based on the transaction serial numbers. Each initial transaction data set contains individual transaction data for a particular flight at a specific point in time. Each transaction data set is processed and classified, and a flight ticket prediction model is trained based on the classified transaction data. After the model is trained, the model is applied to predict flight ticket transactions.

[0117] By using the device provided by the embodiment of the present disclosure, a customer's willingness to purchase air tickets can be predicted based on multiple transaction data, so as to facilitate the subsequent formulation of relevant ticket sales policies.

[0118] In the apparatus provided by the embodiment of the present disclosure, the acquiring unit 401 includes:

[0119] The first acquisition subunit is configured to acquire transaction serial numbers within a preset historical time period;

[0120] a search subunit configured to search the transaction log for an initial transaction data group corresponding to each transaction serial number, wherein the initial transaction data group includes a plurality of initial transaction data;

[0121] The data cleaning subunit is configured to clean each initial transaction data in each of the initial transaction data groups to obtain a plurality of cleaned transaction data groups.

[0122] In the apparatus provided by the embodiment of the present disclosure, the processing unit 402 includes:

[0123] a first determining subunit configured to determine, based on a preset data table, transaction data belonging to direct influencing factors and transaction data belonging to indirect influencing factors in each of the transaction data groups;

[0124] a second acquiring subunit configured to acquire data parameters corresponding to the transaction data of each of the directly influencing factors as characteristic values ​​of the transaction data, and determine that the transaction data of the directly influencing factors are first-category transaction data;

[0125] The processing subunit is configured to split and reorganize the data parameters corresponding to the transaction data belonging to the indirect influencing factors, obtain the characteristic value of each transaction data belonging to the indirect influencing factors, and determine that the transaction data belonging to the indirect influencing factors is the second category of transaction data.

[0126] In the apparatus provided by the embodiment of the present disclosure, the training unit 403 includes:

[0127] a second determining subunit, configured to obtain a preset training target corresponding to each of the transaction data groups, determine a first data weight of each of the first-category transaction data and a second data weight of each of the second-category transaction data in each of the transaction data groups, and generate an actual data weight corresponding to each of the transaction data groups based on each of the first data weights and the second data weights;

[0128] a third determining subunit, configured to use the characteristic value of each of the first-category transaction data and the characteristic value of the second-category transaction data as training data, and the weight of each of the actual data as a model parameter of the air ticket prediction model;

[0129] A training subunit is configured to input each training data group into the air ticket prediction model, triggering the air ticket prediction model to perform iterative training based on the SVM algorithm and each training data;

[0130] A verification subunit, configured to apply the training target corresponding to each transaction data group to verify the training results of each training of the air ticket prediction model;

[0131] The training subunit is stopped and is configured to complete the training of the air ticket prediction model if the current training result meets the training target.

[0132] In the apparatus provided by the embodiment of the present disclosure, the prediction unit 404 includes:

[0133] a third acquisition subunit, configured to acquire a characteristic value of the transaction data to be predicted;

[0134] The prediction subunit is configured to input the characteristic values ​​of the transaction data to be predicted into the air ticket prediction model to obtain the prediction result output by the air ticket prediction model.

[0135] The specific working process of each unit and sub-unit in the air ticket transaction prediction device disclosed in the above embodiment of the present disclosure can be found in the corresponding content of the air ticket transaction prediction method disclosed in the above embodiment of the present disclosure, and will not be repeated here.

[0136] An embodiment of the present disclosure further provides a storage medium, which includes stored instructions, wherein when the instructions are executed, the device where the storage medium is located is controlled to execute the above-mentioned air ticket transaction prediction method.

[0137] The present disclosure also provides an electronic device, as shown in FIG5 . The electronic device includes a memory 501 and one or more instructions 502 . The one or more instructions 502 are stored in the memory 501 and are configured to be executed by one or more processors 503 to perform the following operations:

[0138] Obtaining an air ticket transaction log, and capturing multiple transaction data groups in the transaction log;

[0139] processing each of the transaction data groups to distinguish between first-category transaction data and second-category transaction data in each of the transaction data groups;

[0140] Training an air ticket prediction model based on each of the first-category transaction data and the second-category transaction data;

[0141] When a ticket transaction prediction request corresponding to a target flight is received, the transaction data to be predicted corresponding to the ticket transaction prediction request is obtained, and the transaction data to be predicted is input into the ticket prediction model to obtain the prediction result output by the ticket prediction model, indicating whether all tickets for the target flight have been successfully sold.

[0142] Each embodiment in this specification is described in a progressive manner. The same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments. In particular, for system or system embodiments, since they are basically similar to method embodiments, the description is relatively simple. For relevant parts, refer to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without expending creative work.

[0143] Those skilled in the art may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, computer software, or a combination of both.

[0144] To clearly illustrate the interchangeability of hardware and software, the above descriptions have generally described the components and steps of each example by function. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this disclosure.

[0145] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present disclosure. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure is not limited to the embodiments shown herein, but is intended to be construed in the widest manner consistent with the principles and novel features disclosed herein. Industrial Applicability

[0146] The air ticket transaction prediction method provided by the present disclosure includes: obtaining an air ticket transaction log, capturing multiple transaction data groups in the transaction log; processing each of the transaction data groups, distinguishing between first-category transaction data and second-category transaction data in each of the transaction data groups; training an air ticket prediction model based on each of the first-category transaction data and the second-category transaction data; when an air ticket transaction prediction request corresponding to a target flight is received, obtaining the transaction data to be predicted corresponding to the air ticket transaction prediction request, and inputting the transaction data to be predicted into the air ticket prediction model to obtain a prediction result output by the air ticket prediction model. By applying the method provided by the present disclosure, a customer's willingness to purchase air tickets can be predicted based on multiple transaction data, so as to facilitate the subsequent formulation of relevant ticket sales policies.

Claims

1. A method for predicting air ticket transactions, comprising: Obtaining a ticket transaction log, and capturing a plurality of transaction data groups in the transaction log; Processing each of the transaction data groups to distinguish between first-category transaction data and second-category transaction data in each of the transaction data groups; Based on each of the first-category transaction data and the second-category transaction data, training a ticket prediction model; When a ticket transaction prediction request corresponding to a target flight is received, the transaction data to be predicted corresponding to the ticket transaction prediction request is obtained, and the transaction data to be predicted is input into the ticket prediction model to obtain the prediction result output by the ticket prediction model, that is, whether all tickets for the target flight are successfully sold.

2. The method according to claim 1, wherein: The capturing of multiple transaction data groups in the transaction log includes: Get the transaction serial number within the preset historical time period; Searching the transaction log for an initial transaction data group corresponding to each transaction serial number, wherein the initial transaction data group includes a plurality of initial transaction data; Data cleaning is performed on each of the initial transaction data groups to obtain a plurality of cleaned transaction data groups.

3. The method according to claim 1 or 2, wherein: The processing of each of the transaction data groups to distinguish between the first type of transaction data and the second type of transaction data in each of the transaction data includes: Based on a preset data table, determining transaction data belonging to direct influencing factors and transaction data belonging to indirect influencing factors in each of the transaction data groups; Acquire the data parameters corresponding to the transaction data of each of the direct influencing factors as the characteristic values ​​of the transaction data, and determine that the transaction data of the direct influencing factors are first-category transaction data; Data parameters corresponding to the transaction data belonging to the indirect influencing factors are split and reorganized to obtain characteristic values ​​of each transaction data belonging to the indirect influencing factors, and the transaction data belonging to the indirect influencing factors are determined to be second-category transaction data.

4. The method according to claim 3, wherein: The training of the air ticket prediction model based on each of the first-category transaction data and the second-category transaction data includes: Acquire a preset training target corresponding to each of the transaction data groups, determine a first data weight of each of the first-category transaction data and a second data weight of each of the second-category transaction data in each of the transaction data groups, and generate an actual data weight corresponding to each of the transaction data groups based on each of the first data weights and the second data weights; Using the characteristic value of each of the first-category transaction data and the characteristic value of the second-category transaction data as training data, and using the weight of each of the actual data as a model parameter of the air ticket prediction model; Input each training data group into the air ticket prediction model, triggering the air ticket prediction model to perform iterative training based on the SVM algorithm and each training data; Applying the training target corresponding to each of the transaction data groups to verify the training results of each training of the air ticket prediction model; If the current training result meets the training goal, the training of the air ticket prediction model is completed.

5. The method according to claim 4, wherein: The step of inputting the transaction data to be predicted into the air ticket prediction model to obtain the prediction result output by the air ticket prediction model comprises: Acquiring characteristic values ​​of the transaction data to be predicted; The characteristic values ​​of the transaction data to be predicted are input into the air ticket prediction model to obtain the prediction result output by the air ticket prediction model.

6. An air ticket transaction prediction device, comprising: An acquisition unit, configured to acquire a ticket transaction log and capture a plurality of transaction data groups in the transaction log; A processing unit, configured to process each of the transaction data groups and distinguish between first-category transaction data and second-category transaction data in each of the transaction data groups; A training unit, configured to train an air ticket prediction model based on each of the first-category transaction data and the second-category transaction data; The prediction unit is configured to, when receiving an air ticket transaction prediction request corresponding to a target flight, obtain the transaction data to be predicted corresponding to the air ticket transaction prediction request, and input the transaction data to be predicted into the air ticket prediction model to obtain the prediction result output by the air ticket prediction model, wherein the prediction result is used to provide feedback on whether all the air tickets of the target flight are successfully sold.

7. The device according to claim 6, wherein: The acquisition unit comprises: A first acquisition subunit is configured to acquire a transaction serial number within a preset historical time period; A search subunit is configured to search the transaction log for an initial transaction data group corresponding to each transaction serial number, wherein the initial transaction data group includes a plurality of initial transaction data; The data cleaning subunit is configured to clean each initial transaction data in each of the initial transaction data groups to obtain a plurality of cleaned transaction data groups.

8. The device according to claim 6 or 7, wherein: The processing unit comprises: A first determining subunit is configured to determine, based on a preset data table, transaction data belonging to direct influencing factors and transaction data belonging to indirect influencing factors in each of the transaction data groups; A second acquisition subunit is configured to acquire a data parameter corresponding to the transaction data of each of the directly influencing factors as a characteristic value of the transaction data, and determine that the transaction data of the directly influencing factors is the first type of transaction data; The processing subunit is configured to split and reorganize the data parameters corresponding to the transaction data belonging to the indirect influencing factors, obtain the characteristic value of each transaction data belonging to the indirect influencing factors, and determine that the transaction data belonging to the indirect influencing factors is the second category of transaction data.

9. The device according to claim 8, wherein: The training unit comprises: a second determining subunit, configured to obtain a preset training target corresponding to each of the transaction data groups, and determine a first data weight of each of the first-category transaction data and a second data weight of each of the second-category transaction data in each of the transaction data groups, and generate an actual data weight corresponding to each of the transaction data groups based on each of the first data weights and the second data weights; A third determination subunit is configured to use the characteristic value of each of the first-category transaction data and the characteristic value of the second-category transaction data as training data, and each of the actual data weights as a model parameter of the air ticket prediction model; A training subunit, configured to input each training data group into the air ticket prediction model, triggering the air ticket prediction model to perform iterative training based on the SVM algorithm and each training data; A verification subunit, configured to apply the training target corresponding to each of the transaction data groups to verify the training result of each training of the air ticket prediction model; The training subunit is stopped and is set to complete the training of the air ticket prediction model if the current training result meets the training target.

10. The device according to claim 9, wherein: The prediction unit comprises: A third acquisition subunit is configured to acquire a characteristic value of the transaction data to be predicted; The prediction subunit is configured to input the feature value of the transaction data to be predicted into the air ticket prediction model to obtain the prediction result output by the air ticket prediction model.

Images