Dynamic adjustment method, system, device and storage medium for air ticket reservation

Abstract

The present invention provides a dynamic adjustment method, system, device and storage medium for air ticket reservation. The method includes: extracting features of air ticket data through a decision tree, pruning the subdivided nodes to obtain N features that affect availability and booking as important features, where N is a preset value; obtaining prohibited entities based on the historical data of the previous time window through an algorithm model; obtaining the prohibited duration for each prohibited entity based on the number of prohibited times in several consecutive past time windows; when the pass rate pass of the prohibited entity is less than the prohibition threshold, it enters the prohibition; otherwise, it does not enter the prohibition; obtaining unique prohibited entities through a deduplication and merging algorithm. The present invention can timely discover the ticketing management defects of airlines and make timely production dynamic adjustments, improving the process experience of users booking air tickets.

Description

Background Art

[0002] When a user books an international flight ticket, the pre-booking process mainly goes through five pages: a query page for specifying query conditions, a filling page for entering passenger information, a value-added page for purchasing additional products or services, a payment page for paying fees, and a completion page for confirming order details. After the user specifies query parameters such as itinerary type, departure and destination, date, cabin class, passenger type, etc. on the query page, the system initiates a search for the user to each quotation system, merges the results and displays them to the user. After the user browses the quotation list, considering multiple factors such as route, quotation, refund and change, luggage, auxiliary products, etc., the user selects a favorite flight group and starts the reservation process. On each page, the background will initiate some key validations to help the user better confirm the itinerary. For example, when the user enters the filling page, the background will initiate a bookability validation, which includes two parts: fare validation and availability validation. Fare validation is used to verify whether the fare calculation and tax item calculation are correct when the engine quotes, and availability validation is used to verify whether the seat availability is accurate when quoting. Another example is that when the user is on the value-added page, the background will initiate a seat reservation operation and send a request to reserve seats to the Global Distribution System (GDS). The GDS will simultaneously send a request to the airline's Central Reservation System (CRS). During this process, the airline conducts various information validations on passengers, itinerary, etc., and the most important one is the validation of the consistency between the seat data of the GDS and the airline's seats. If these key validations fail, there will be front-end interception prompts such as "price unavailable" and "seats sold out", and the reservation experience is not smooth, causing dissatisfaction among users. It can be seen that a reliable quotation system needs to continuously output high-quality and high-stability quotations under the dual constraints of cost and coverage rate.

[0003] Therefore, the present invention provides a dynamic adjustment method, system, device and storage medium for air ticket reservation. Summary of the Invention

[0004] Aiming at the problems in the prior art, the purpose of the present invention is to provide a dynamic adjustment method, system, device and storage medium for air ticket reservation, which overcomes the difficulties of the prior art, can timely discover the defects in the airline's ticket management and make production dynamic adjustments in a timely manner, and improves the process experience of users booking air tickets.

[0005] An embodiment of the present invention provides a dynamic adjustment method for air ticket reservation, including the following steps:

[0006] S110. Extract features of air ticket data through a decision tree, prune the subdivided nodes to obtain N features affecting bookability and seat reservation as important features, where N is a preset value;

[0007] S120. Obtain prohibited entities based on the historical data of the previous time window and the algorithm model;

[0008] S130. Obtain the prohibited duration for each prohibited entity based on the number of times of prohibition in several consecutive past time windows;

[0009] S140. When the pass rate pass of the prohibited entity is less than the prohibition threshold, enter the prohibition; otherwise, do not enter the prohibition;

[0010] S150. Obtain the unique prohibited entities through the algorithm of deduplication and merging.

[0011] Preferably, in the step S110, the important features include at least one of itinerary type, departure city, departure country, cabin sales location, reservation data warehouse, and ticket issuing counter.

[0012] Preferably, in the step S120, the following steps are included:

[0013] S121. The algorithm learns the historical data of the previous time window, constructs a model, and applies it to the current time window;

[0014] S122. Select the time window t, each time window has the same length, and the length is r;

[0015] S123. Select a stage f, the bookable stage or the reservation stage, based on N feature dimensions as data columns, and the data of this stage within the preset time range as data rows M, to form a data matrix and train the model;

[0016] S124. Loop i = 0:N - 1, and execute each round of leveli:

[0017] S125. In the i-th round, select i features. For each feature, based on the above data matrix, aggregate N feature dimensions, calculate the total amount and the pass rate respectively, and generate prohibited entities composed of (N - i) explicit features and i virtual features.

[0018] Preferably, in the step S122, as the length r of the time window decreases, the accuracy of the model increases;

[0019] In the step S124, it includes: the i = 0 round is the finest granularity. When the pass rate is lower than the set threshold under the features of the i = 0 round, add a prohibition mark to the data object.

[0020] Preferably, in the step S130, the following steps are included:

[0021] S131. Obtain the embargo status of each of the said entities subject to embargo in the past T consecutive time windows. Define the status st = 0, indicating no embargo; st = 1, indicating an embargo. The value range of t is from 1 to T;

[0022] S132. Define a penalty factor Calculate the number of times the entity has been marked as subject to embargo in the historical window period;

[0023] S133. Define a reward factor It represents the proportion of the longest continuous length during which the entity has been continuously marked as not subject to embargo in the historical window period. Among them, L represents the maximum continuous length from t = 1 when st = 0 is possible in the historical time window;

[0024] S134. Obtain the embargo duration d = (1 - pass) * r * (1 + max{p - b, 0}).

[0025] Preferably, in the said step S140, the following steps are included:

[0026] S141. Based on the intersection of the true situation and the model prediction situation in the confusion matrix, obtain four situations: true negative, false negative, false positive, and true positive. Among them, negative corresponds to the model not implementing an embargo, positive corresponds to the model implementing an embargo, true corresponds to the model prediction result being consistent with the actual situation, and false corresponds to the model prediction result being inconsistent with the actual situation;

[0027] S142. Define the improvement rate as the proportion of those actually intercepted and subject to embargo, that is, TN / All;

[0028] S143. Define the mis - embargo rate as the proportion of those actually passing through but subject to embargo, that is, FN / All;

[0029] S144. Traverse all embargo thresholds according to the preset step interval, and calculate the curve of the improvement rate and the mis - embargo rate under the thresholds;

[0030] S145. Calculate the marginal utility of improvement and the marginal cost of mis - embargo, and find the embargo threshold with a large improvement rate and a small mis - embargo rate, which is considered the most valuable embargo threshold;

[0031] S146. Take the embargo threshold with a large improvement rate and a small mis - embargo rate as the most fine - grained embargo threshold, and the embargo thresholds at the remaining each Level are attenuated at a uniform step of the embargo threshold with a large improvement rate and a small mis - embargo rate / N.

[0032] Preferably, in the said step S150, when the high - Level embargo message contains the low - Level embargo message, remove the low - Level embargo message and only retain the high - Level embargo message.

[0033] An embodiment of the present invention also provides a dynamic adjustment system for air ticket reservation, which is used to implement the above-mentioned dynamic adjustment method for air ticket reservation. The dynamic adjustment system for air ticket reservation includes:

[0034] A feature extraction module that extracts features of air ticket data through a decision tree, prunes the subdivided nodes to obtain N features that affect availability and reservation as important features, where N is a preset value;

[0035] A prohibited sale entity module that obtains prohibited sale entities based on an algorithm model based on historical data of the previous time window;

[0036] A prohibited sale duration module that obtains the prohibited sale duration for each prohibited sale entity based on the number of prohibited sale times in several consecutive past time windows;

[0037] An execution of prohibited sale module that enters prohibited sale when the pass rate pass of the prohibited sale entity is less than the prohibited sale threshold; otherwise, it does not enter prohibited sale;

[0038] A duplicate removal and merging module that obtains unique prohibited sale entities through a duplicate removal and merging algorithm.

[0039] An embodiment of the present invention also provides a dynamic adjustment device for air ticket reservation, including:

[0040] A processor;

[0041] A memory that stores executable instructions of the processor;

[0042] Wherein, the processor is configured to execute the steps of the above-mentioned dynamic adjustment method for air ticket reservation by executing the executable instructions.

[0043] An embodiment of the present invention also provides a computer-readable storage medium for storing a program, and when the program is executed, the steps of the above-mentioned dynamic adjustment method for air ticket reservation are implemented.

[0044] The purpose of the present invention is to provide a dynamic adjustment method, system, device and storage medium for air ticket reservation, which can timely discover the defects in the ticket management of airlines and make timely production dynamic adjustments, improving the process experience of users booking air tickets. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] By reading the detailed description of the non-limiting embodiments with reference to the following drawings, other features, objects and advantages of the present invention will become more apparent.

[0046] Figure 1 is a flowchart of the dynamic adjustment method for air ticket reservation of the present invention.

[0047] Figure 2 is a schematic diagram of an aggressive model in the implementation process of the dynamic adjustment method for air ticket reservation of the present invention.

[0048] Figure 3 It is a schematic diagram of the conservative model in the implementation process of the dynamic adjustment method for air ticket reservation of the present invention.

[0049] Figure 4 It is a schematic diagram of the compromise model in the implementation process of the dynamic adjustment method for air ticket reservation of the present invention.

[0050] Figure 5 It is a schematic diagram of the modules of the dynamic adjustment system for air ticket reservation of the present invention.

[0051] Figure 6 It is a schematic diagram of the structure of the dynamic adjustment device for air ticket reservation of the present invention.

[0052] Figure 7 It is a schematic diagram of the structure of the computer-readable storage medium according to an embodiment of the present invention. Detailed implementation manners

[0053] The following uses specific specific examples to illustrate the implementation manners of the present application. Those skilled in the art can easily understand other advantages and effects of the present application from the content disclosed in the present application. The present application can also be implemented or applied through other different specific implementation manners. Various details in the present application can also be modified or changed according to different viewpoints and application systems without departing from the spirit of the present application. It should be noted that, without conflict, the embodiments and features in the embodiments of the present application can be combined with each other.

[0054] The following takes the accompanying drawings as a reference and details the embodiments of the present application so that those skilled in the technical field to which the present application belongs can easily implement it. The present application can be embodied in many different forms and is not limited to the embodiments described herein.

[0055] In the description of the present application, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. mean that the specific features, structures, materials, or characteristics represented in connection with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials, or characteristics represented can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples represented in the present application and the features of the different embodiments or examples.

[0056] In addition, the terms "first" and "second" are used for illustrative purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In the representation of this application, "a plurality of" means two or more, unless otherwise specifically defined.

[0057] To clearly illustrate this application, devices irrelevant to the description are omitted, and the same or similar constituent elements throughout the specification are given the same reference numerals.

[0058] Throughout the specification, when it is said that a device is "connected" to another device, this includes not only the case of "direct connection" but also the case of "indirect connection" with other elements interposed therebetween. Additionally, when it is said that a certain device "includes" a certain constituent element, unless there is a particularly contrary record, it does not exclude other constituent elements but means that other constituent elements may also be included.

[0059] When it is said that a device is "above" another device, this may be directly above the other device, but there may also be other devices in between. When it is said that a device is "directly" "above" another device, there are no other devices in between.

[0060] Although in some instances the terms first, second, etc. are used herein to denote various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, the first interface and the second interface, etc. are indicated. Furthermore, as used herein, the singular forms "a", "an", and "the" are also intended to include the plural forms unless the context indicates otherwise. It should be further understood that the terms "comprising", "including" indicate the presence of the features, steps, operations, elements, components, items, kinds, and / or groups, but do not exclude the presence, occurrence, or addition of one or more other features, steps, operations, elements, components, items, kinds, and / or groups. The terms "or" and "and / or" used herein are interpreted as inclusive, or meaning any one or any combination. Thus, "A, B, or C" or "A, B, and / or C" means "any one of the following: A; B; C; A and B; A and C; B and C; A, B, and C". An exception to this definition occurs only when the combination of elements, functions, steps, or operations is inherently mutually exclusive in some way.

[0061] The technical terms used herein are for the purpose of referring to specific embodiments only and are not intended to limit the present application. The singular forms used herein also include the plural forms as long as the context does not clearly indicate the contrary meaning. The meaning of "including" used in the specification is to specify particular features, regions, integers, steps, operations, elements, and / or components, and does not exclude the existence or addition of other features, regions, integers, steps, operations, elements, and / or components.

[0062] Although not defined differently, all terms, including the technical terms and scientific terms used herein, have the same meaning as generally understood by those skilled in the technical field to which the present application pertains. Terms defined in commonly used dictionaries are additionally interpreted to have meanings consistent with the relevant technical literature and the content presented herein. As long as they are not defined, they should not be over-interpreted as ideal or overly formulaic meanings.

[0063] Currently, fare control can generally be divided into freight rate control and seat inventory control. Freight rate control is mainly reflected in the differentiated initial pricing under different condition ranges, such as different prices under different periods (off-peak and peak seasons, mid-week and weekends), different refund and change conditions, and different baggage conditions. Therefore, freight rate data is relatively static. At the same time, the cost for Trip to obtain freight rate data is fixed, and the accuracy is not restricted by cost. As the data manifestation of the airline's revenue management (or yield management) system, seat inventory essentially depends on the regulation of supply and demand. When a seat for a certain fare is booked, the airline's revenue management system will recalculate and decide whether to continue to open that seat. If user behavior can be regarded as a stochastic process, then seat inventory data will be constantly randomly perturbed by user booking behavior and keep changing. In addition, users always tend to choose low-price products among all quotes, and low-price quotes often have large and rapid demand changes. After the system has been iteratively optimized until now, whether it is seat inventory or booking, currently user booking interception is mainly restricted by the unpredictability of external data. The in-house engine covers more than 700 global airlines, covering the vast majority of full-service airlines and some low-cost airlines. The source of seat inventory data is each GDS or airline, and the accuracy of the original data provided to Trip may have random problems to varying degrees in terms of airlines, routes, regions, etc., manifested as local and occasional data problems or business problems. After a series of revenue management-related optimizations were launched for seat inventory and booking, it became possible to run a "pass rate feedback" data model based on the big data of user booking behavior. It can be seen that the factors affecting user interception are basically external factors, and they are long-tailed and cannot be predicted in advance. Currently, the practices of each quote engine are relatively mechanical. Either after a single interception occurs, the product is prohibited from being sold for a fixed number of hours according to a series of attributes; or after multiple interceptions occur, products with low quality are prohibited from being sold manually for a long time, and the prohibition is reviewed and lifted regularly. And currently, all the prohibitions occur after the user has "already" encountered interception, lacking foresight.

[0064] Figure 1 is a flowchart of the dynamic adjustment method for air ticket reservation of the present invention. As Figure 1 shown, an embodiment of the present invention provides a dynamic adjustment method for air ticket reservation, including the following steps:

[0065] S110. Extract features of air ticket data through a decision tree, prune the subdivided nodes to obtain N features affecting availability and reservation as important features, where N is a preset value;

[0066] S120. Obtain prohibited entities based on the historical data of the previous time window based on an algorithm model;

[0067] S130. For each prohibited entity, obtain the prohibited duration based on the number of prohibited times in several consecutive past time windows;

[0068] S140. When the pass rate pass of the prohibited entity is less than the prohibition threshold, enter the prohibition; otherwise, do not enter the prohibition;

[0069] S150. Obtain unique prohibited entities through a deduplication and merging algorithm.

[0070] In a preferred embodiment, in step S110, the important features include at least one of itinerary type, departure city, departure country, cabin sales location, reservation data warehouse, and ticket issuing counter.

[0071] In a preferred embodiment, in step S120, it includes the following steps:

[0072] S121. The algorithm learns the historical data of the previous time window, constructs a model, and applies it to the current time window;

[0073] S122. Select a time window t, each time window has the same length, and the length is r;

[0074] S123. Select a stage f, the availability stage or the reservation stage, based on N feature dimensions as data columns, and the data of this stage within a preset time range as data rows M to form a data matrix and train the model;

[0075] S124. Loop i = 0:N-1, and execute each round of level i:

[0076] S125. In the i-th round, select i features. For each feature, based on the above data matrix, aggregate N feature dimensions, calculate the total amount and the pass rate respectively, and generate a prohibited entity composed of (N-i) explicit features and i virtual features

[0077] In a preferred embodiment, in step S122, as the length r of the time window decreases, the accuracy of the model increases;

[0078] In step S124, it includes: the i = 0 round is the finest granularity. When the passing rate is lower than the set threshold under the features of this round, a sales ban mark is added to the data object.

[0079] In a preferred embodiment, in step S130, it includes the following steps:

[0080] S131. Obtain the sales ban status of each entity subject to the sales ban in the past T consecutive time windows. Define the status st = 0, then it is not subject to the sales ban; st = 1, then it is subject to the sales ban, where the value range of t is from 1 to T;

[0081] S132. Define the penalty factor Calculate the number of times the entity has been marked as subject to the sales ban in the historical window period;

[0082] S133. Define the reward factor It represents the proportion of the longest continuous length that the entity has been continuously marked as not subject to the sales ban in the historical window period, where L represents the maximum continuous length from t = 1 when st = 0 in the historical time window;

[0083] S134. Obtain the sales ban duration d = (1 - pass) * r * (1 + max{p - b, 0}).

[0084] In a preferred embodiment, in step S140, it includes the following steps:

[0085] S141. Based on the intersection of the true situation and the model prediction situation in the confusion matrix, obtain four situations: true negative, false negative, false positive, and true positive. Among them, negative corresponds to the model not implementing the sales ban, positive corresponds to the model implementing the sales ban, true corresponds to the model prediction result being consistent with the actual situation, and false corresponds to the model prediction result being inconsistent with the actual situation;

[0086] S142. Define the improvement rate as the proportion of those actually intercepted and subject to the sales ban, that is, TN / All,

[0087] S143. Define the false sales ban rate as the proportion of those actually passing but subject to the sales ban, that is, FN / All,

[0088] S144. Traverse all sales ban thresholds according to the preset step interval, and calculate the curve of the improvement rate and the false sales ban rate under the thresholds.

[0089] S145. Calculate the marginal utility of improvement and the marginal cost of false sales ban, and find the sales ban threshold with a large improvement rate and a small false sales ban rate, which is considered the most valuable sales ban threshold.

[0090] S146. Use this value as the finest-grained sales ban threshold, and the sales ban thresholds for the remaining levels decay uniformly at a step of this value / N.

[0091] In a preferred embodiment, in step S150, when the high-level sales ban message contains the low-level sales ban message, remove the low-level sales ban message and only retain the high-level sales ban message.

[0092] Through the constructed sales ban system, the present invention can actively intervene in advance when it is known that the predetermined quality is not ideal. In addition, the statically matched sales ban entities and the sales ban duration with a fixed period also need to be optimized into a dynamic manner, and it is hoped that the model has the ability of "intelligent" abstraction. The model system obtained by machine learning can ban a series of entities with common problems in advance, preventing users from encountering them again and directly improving the predetermined quality. More importantly, after consuming the model, the quotation engine can output a sub-low quotation with better quality and better conversion for users to choose, rather than simply subtracting on the result set.

[0093] In recent years, with the improvement of technology, business expansion and globalization strategy, the quotation resources for Trip international tickets have gradually become rich, the route coverage has been continuously expanded, and the competitive advantage has been continuously growing. At the same time, the smoothness of users at each reservation stage, as part of the system reliability, is also crucial and needs to be continuously improved and optimized. This dynamic model is committed to creating a highly smooth reservation system, which on the one hand helps to promote conversion and increase revenue, and on the other hand also conforms to the high-quality service concept promised by Trip.

[0094] The specific implementation process of the present invention is as follows:

[0095] (1) Data preparation.

[0096] Use the bookable data of users on the filling page and the seat reservation data of users on the value-added page.

[0097] (2) Feature selection.

[0098] Use the CART decision tree for feature extraction, prune the overly subdivided nodes, and obtain N important features that affect bookability and seat reservation. Further, these features extracted based on the classification tree method also need to be deeply consistent with the understanding of the international ticket business. For example, the trip type, origin city, origin country are closely related to the POC revenue management method adopted by European airlines; the POS revenue management method adopted by North American airlines is highly related to the point of sale (POS), booking GDS, and ticketing agency. Information about city pairs and country pairs helps to capture randomly scattered route problems.

[0099] (3) Algorithm Design

[0100] (3.1) Prohibited Sale Entities

[0101] Prohibited sale entities refer to a type of data object used for prohibited sales.

[0102] The algorithm learns the historical data of the previous time window, constructs a model, and applies it to the current time window.

[0103] Select time window time t. Each time window has the same length, which is r. The model can be refined by continuously reducing r, which is equivalent to differentiation. The value of r needs to consider the trade-off between refinement and overfitting.

[0104] Select a phase phase f, which can be the order placement phase or the reservation phase. Based on N feature dimensions as data columns, and the data of this phase within a certain time range as data rows M, a data matrix is formed to train the model. The key metrics used here are the total quantity (counts) and the passing rate (pass).

[0105] Loop i = 0:N-1, and execute each round leveli:

[0106] In the i-th round, select i features. For each feature, based on the above data matrix, aggregate N feature dimensions, calculate the total quantity and the passing rate respectively, and generate prohibited sale entities composed of (N-i) explicit features and i virtual features. Among them, the 0-th round is the finest granularity, and the entity has no virtual features. If the passing rate is lower than the set threshold under this round of features, a prohibited sale mark is made.

[0107] (3.2) Prohibited Sale Duration

[0108] Linearly related to the passing rate formula, calculate the current prohibited sale duration. Based on the data performance of multiple historical time windows, a reward and punishment mechanism is set for the prohibited sale duration. For products with excellent recent performance in history, when the current product quality is slightly worse, the prohibited sale duration is set relatively shorter (reward); for products with poor long-term performance in history, the prohibited sale duration is set more severely (punishment). The reward and punishment mechanism makes the prohibited sale duration more reasonable.

[0109] For a prohibited sale entity,

[0110] The historical time window is composed of a series of consecutive past time windows, from near to far, t ranges from 1 to T.

[0111] Define the status State s t = 0, then there is no prohibited sale;

[0112] s t = 1, then there is a prohibited sale.

[0113] Define the penalty factor Calculate the number of times the entity is marked as prohibited from sale during the historical window period; it represents the penalty intensity for repeated interception.

[0114] Define the reward factor It represents the proportion of the longest length that the entity is continuously marked as not prohibited from sale during the historical window period. Among them, L represents the historical time window starting from t = 1, and s t = 0 is the maximum continuous length, which is used to measure the number of times with better "continuous" quality. The proportion of the number of continuously better times to the size of the historical window is used as the reward intensity.

[0115] Calculate the duration of prohibited sale

[0116] Duration d = (1 - pass) * r * (1 + max{p - b, 0})

[0117] (3.3) Prohibited sale threshold

[0118] Given a prohibited sale threshold, if the passing rate of the prohibited entity is less than the prohibited sale threshold, it enters the prohibited sale; otherwise, it does not enter the prohibited sale. Based on the idea of the Confusion Matrix, the true situation and the model prediction situation intersect to obtain four possibilities: true negative (TN, True Negative), false negative (FN, False Negative), false positive (FP, False Positive), and true positive (TP, True Positive). Yin and yang respectively correspond to the model not implementing the prohibited sale and implementing the prohibited sale, and true and false respectively correspond to the model prediction result being consistent and inconsistent with the actual situation. Define the improvement rate as the proportion of the actually intercepted and entering the prohibited sale, that is, TN / All, which reflects the improvement in the passing rate in production after the prohibited sale model takes effect. Define the false prohibition rate as the proportion of the actually passed but entering the prohibited sale, that is, FN / All, which can be regarded as the opportunity cost brought by the prohibited sale. Traverse all possible prohibited sale thresholds at a certain step interval gap g, and calculate the curves of the improvement rate and the false prohibition rate under the thresholds. Calculate the marginal utility of improvement and the marginal cost of false prohibition, and find the prohibited sale threshold with a large improvement rate and a small false prohibition rate, which is considered the most valuable prohibited sale threshold. Take this value as the finest-grained (Level 0) prohibited sale threshold, and the prohibited sale thresholds of the remaining levels are attenuated at a uniform step of this value / N.

[0119] (3.4) Deduplication and merging

[0120] In actual implementation, the operations at level 0 are based on the original data. In theory, the data at level 1... level N should be based on the data at level N - 1. For the sake of simplifying the calculation, they are all based on level 0. This makes there be an inclusion relationship among all prohibited entities, and this relationship is redundant. The same prohibited record may contribute to different prohibited entities. A deduplication and merging algorithm is needed to show clear and unique prohibited entities. When a high-level prohibited message contains a low-level prohibited message, only the high-level prohibited message is retained.

[0121] Four deduplication and merging algorithms are designed. Essentially, in the constructed multi-way tree, find the path to the level where a certain prohibited entity is located.

[0122] The first aggressive model ( Figure 2 ) allows cross-level merging at any level and is relatively free. The nodes in the figure represent prohibited entities and their levels, and the connections indicate an inclusion relationship between two entities, where the high-level entity includes the low-level one. The second conservative model ( Figure 3 ) does not allow any cross-level merging, that is, all retained prohibited entities satisfy that for each level below the level where the entity is located, there is a corresponding prohibited entity. That is to say, the depth of the subtree must be equal to level + 1. The dotted nodes in the figure represent the prohibited entities that cannot be generated compared with the aggressive model. The dotted nodes do not actually exist and are only used to illustrate that the reason why L3 cannot be generated is the lack of support from L1 and L2.

[0123] The third compromise model ( Figure 4 ) does not allow arbitrary cross-level merging and only allows downward merging when there are adjacent-level entities. Different from the conservative model, the prohibited entities retained by the compromise model only need to satisfy that there is a path (not all paths) such that the tree depth is less than or equal to level + 1. The lower the level of the prohibited entity, the finer and more scattered the granularity; the higher the level, the greater the improvement in prohibition, and the merging needs to be extremely cautious. Relatively speaking, the compromise model combines the aggressive model and the conservative model and has the best balance.

[0124] However, in some scenarios, the compromise model still inevitably generates prohibited entities at low levels, weakening the generalization ability of the model; then a composite version is optimized on the basis of the compromise model. Select n, and apply the compromise model between levels n + 1... N; apply the aggressive model between levels 0 - n. Iteratively optimize n to find the optimal composite model.

[0125] (4) Model evaluation

[0126] In addition to using the improvement rate and false ban rate defined in (3.3) to evaluate the business effect, the classic general accuracy (Precision), recall rate (Recall), and accuracy metrics of machine learning are also used to evaluate the model.

[0127] (5) Model application

[0128] The model is connected to the message management center as a producer. When a new embargo message is generated, it will call the write interface provided by the message management center to write the latest message. The international flight ticket engine acts as a consumer and listens to the message management center in real time. When there is a new message, it will immediately consume it and output a sub-optimal solution.

[0129] In addition, the model can help discover production problems with clinical significance, which are all helpful for making timely production responses. It can exactly help discover the revenue management logic behind the phenomena of "searchable but not bookable" and "phantom cabins" of airlines, or third-party data problems caused by inconsistent GDS seats and airline seats.

[0130] Figure 5 It is a schematic diagram of the modules of the dynamic adjustment system for flight ticket reservation of the present invention. As Figure 5 shown, the dynamic adjustment system 5 for flight ticket reservation of the present invention includes:

[0131] A feature extraction module 51 extracts features of flight ticket data through a decision tree, prunes the subdivided nodes to obtain N features that affect bookability and seat reservation as important features, where N is a preset value;

[0132] An embargo entity module 52 obtains embargo entities based on algorithm models from historical data of the previous time window;

[0133] An embargo duration module 53 obtains the embargo duration for each embargo entity based on the number of embargoes in several consecutive past time windows;

[0134] An embargo execution module 54 enters the embargo when the pass rate pass of the embargo entity is less than the embargo threshold; otherwise, it does not enter the embargo;

[0135] A duplicate removal and merging module 55 obtains unique embargo entities through duplicate removal and merging algorithms.

[0136] The dynamic adjustment system for flight ticket reservation of the present invention can timely discover defects in the airline's ticket management and make timely production dynamic adjustments, improving the process experience of users booking flight tickets.

[0137] The above embodiments are only preferred examples of the present invention and are not used to limit the present invention. Any equivalent substitutions, modifications, and changes made within the principles of the present invention are within the protection scope of the present invention.

[0138] An embodiment of the present invention further provides a dynamic adjustment device for ticket reservation, including a processor and a memory, in which executable instructions of the processor are stored. The processor is configured to execute the steps of the dynamic adjustment method for ticket reservation by executing the executable instructions.

[0139] As shown above, the dynamic adjustment system for ticket reservation of the present invention can timely detect the defects in the ticket management of airlines and make timely production dynamic adjustments, improving the process experience of users booking tickets.

[0140] Those skilled in the art can understand that various aspects of the present invention can be implemented as a system, a method, or a program product. Therefore, various aspects of the present invention can be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or an implementation combining hardware and software aspects, which can be collectively referred to as "circuit", "module", or "platform" here.

[0141] Figure 6 is a schematic structural diagram of the dynamic adjustment device for ticket reservation of the present invention. The following will refer to Figure 6 to describe the electronic device 600 according to this embodiment of the present invention. Figure 6 The electronic device 600 shown is only an example and should not bring any limitation to the functions and usage scope of the embodiments of the present invention.

[0142] As Figure 6 shown, the electronic device 600 is presented in the form of a general computing device. The components of the electronic device 600 may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different platform components (including the storage unit 620 and the processing unit 610), a display unit 640, etc.

[0143] Among them, the storage unit stores program codes, and the program codes can be executed by the processing unit 610, so that the processing unit 610 executes the steps according to various exemplary embodiments of the present invention described in the above-mentioned electronic prescription transfer processing method part of this specification. For example, the processing unit 610 can execute the steps as Figure 1 shown in

[0144] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 6201 and / or a cache storage unit 6202, and may further include a read-only storage unit (ROM) 6203.

[0145] The storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205. Such program modules 6205 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment.

[0146] The bus 630 may represent one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus structures.

[0147] The electronic device 600 may also communicate with one or more external devices 700 (such as a keyboard, a pointing device, a Bluetooth device, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 600, and / or may communicate with any device that enables the electronic device 600 to communicate with one or more other computing devices (such as a router, a modem, etc.). Such communication may be through the input / output (I / O) interface 650. Also, the electronic device 600 may communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through the network adapter 660. The network adapter 660 may communicate with other modules of the electronic device 600 through the bus 630. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms, etc.

[0148] An embodiment of the present invention also provides a computer-readable storage medium for storing a program, and the steps of a dynamic adjustment method for air ticket reservation are implemented when the program is executed. In some possible implementation manners, various aspects of the present invention may also be implemented in the form of a program product, which includes program code. When the program product runs on a terminal device, the program code is used to cause the terminal device to execute the steps according to various exemplary embodiments of the present invention described in the above-mentioned electronic prescription circulation processing method part of this specification.

[0149] As shown above, the dynamic adjustment system for air ticket reservation according to this embodiment of the present invention can timely discover the ticketing management defects of airlines and make timely production dynamic adjustments, improving the process experience of users booking air tickets.

[0150] Figure 7 It is a schematic structural diagram of the computer-readable storage medium of the present invention. Refer to Figure 7As shown, a program product 800 for implementing the above method according to an embodiment of the present invention is described. It can be a portable compact disc read-only memory (CD-ROM) and includes program code, and can run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto. In this document, a readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0151] The program product can adopt any combination of one or more readable media. The readable media can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0152] The computer-readable storage medium can include a data signal propagated in a baseband or as part of a carrier wave, in which the readable program code is carried. Such a propagated data signal can take various forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The readable storage medium can also be any readable medium other than the readable storage medium that can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium can be transmitted with any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.

[0153] The program code for performing the operations of the present invention can be written in any combination of one or more programming languages. The programming languages include object-oriented programming languages - such as Java, C++, etc., and also include conventional procedural programming languages - such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, executed as an independent software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user's computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, by using an Internet service provider to connect through the Internet).

[0154] In summary, the purpose of the present invention is to provide a method, system, device and storage medium for dynamically adjusting flight ticket reservations, which can timely detect the defects in the airline's ticket management and make timely production dynamic adjustments, so as to improve the process experience of users booking flight tickets.

[0155] The above content is a further detailed description of the present invention in combination with specific preferred implementation manners. It cannot be determined that the specific implementation of the present invention is only limited to these descriptions. For those of ordinary skill in the technical field to which the present invention belongs, without departing from the concept of the present invention, several simple deductions or replacements can still be made, and all should be regarded as belonging to the protection scope of the present invention.

Claims

1. A dynamic adjustment method for air ticket reservation, characterized in that Including the following steps: S110. Extract features of ticket data through a decision tree, prune the subdivided nodes to obtain N features that affect bookability and reservation as important features, where N is a preset value; S120. Learn the historical data of the previous time window of the algorithm, construct a model, and apply it to the current time window; select a time window t, each time window has the same length, and the length is r; select a stage f, which can be the ordering stage or the positioning stage. Based on N feature dimensions as data columns, the data of this stage within the preset time range is used as data rows M to form a data matrix and train the model; loop i = 0:N - 1, and execute each round of level i: in the i-th round, select i features. For features, based on the above data matrix, aggregate N feature dimensions, calculate the total amount and the passing rate respectively, and generate a prohibited entity composed of (N - i) explicit features and i virtualized features; S130. Obtain the prohibited status of each of the said prohibited entities in the past T consecutive time windows, and define the status S t = 0, then it is not prohibited from being sold; S t = 1, then it is prohibited from being sold, and the value range of t is from 1 to T; Define the penalty factor Calculate the number of times the entity has been marked as prohibited during the historical window period; Define the reward factor represents the proportion of the longest length that the entity has been continuously marked as not prohibited during the historical window period, where L represents the historical time window starting from t = 1, and S t = 0 is the maximum continuous length; Obtain the prohibited duration d = (1 - pass)*r*(1 + max{p - b, 0}), where pass is the passing rate of the prohibited entity S140. Based on the intersection of the actual situation and the model prediction situation in the confusion matrix, four situations of true negative, false negative, false positive, and true positive are obtained. Among them, negative corresponds to the model not performing a sales ban, positive corresponds to the model performing a sales ban, true corresponds to the model prediction result being consistent with the actual situation, and false corresponds to the model prediction result being inconsistent with the actual situation; define the improvement rate as the proportion of actual interceptions that enter the sales ban, that is, TN / All; define the misban rate as the proportion of those that actually pass but enter the sales ban, that is, FN / All; traverse all sales ban thresholds according to the preset step interval, and calculate the curve of the improvement rate and misban rate under the threshold; calculate the marginal utility of improvement and the marginal cost of misban, and find the sales ban threshold with a large improvement rate and a small misban rate, which is considered the most valuable sales ban threshold; use the sales ban threshold with a large improvement rate and a small misban rate as the finest-grained sales ban threshold, and the sales ban thresholds at the remaining levels decay uniformly at a step of the sales ban threshold with a large improvement rate and a small misban rate / N; S150. Obtain unique sales ban entities through a deduplication and merging algorithm.

2. The dynamic adjustment method for air ticket reservation according to claim 1, wherein In the step S110, the important features include at least one of itinerary type, departure city, departure country, cabin sales location, reservation data warehouse, and ticket issuance counter.

3. The dynamic adjustment method for air ticket reservation according to claim 1, characterized in that In the step S120, as the length r of the time window decreases, the accuracy of the model increases; In the step S120, it includes: the i = 0 round is the finest-grained. When the passing rate is lower than the set threshold under the features of the i = 0 round, a sales ban mark is added to the data object.

4. The dynamic adjustment method for air ticket reservation according to claim 1, characterized in that In the step S150, when the high-level sales ban message contains the low-level sales ban message, the low-level sales ban message is removed, and only the high-level sales ban message is retained.

5. A dynamic adjustment system for air ticket reservation, which is used to implement the dynamic adjustment method for air ticket reservation described in claim 1, and is characterized in that Including: A feature extraction module that extracts features of ticket data through a decision tree, prunes the subdivided nodes to obtain N features that affect bookability and reservation as important features, where N is a preset value; The prohibited sales entity module learns the historical data of the previous time window through algorithms, constructs a model, and applies it to the current time window; select a time window t, each time window has the same length, and the length is r; select a stage f, which can be the booking stage or the seat reservation stage. Based on N feature dimensions as data columns, the data of this stage within a preset time range is used as data rows M to form a data matrix and train the model; loop i = 0:N-1, and execute each round of level i: in the i-th round, select i features. For features, based on the above data matrix, aggregate N feature dimensions, calculate the total quantity and the passing rate respectively, and generate a prohibited sales entity composed of (N-i) explicit features and i virtual features. The prohibited sales duration module obtains the prohibited sales status of each of the said prohibited sales entities in the past T consecutive time windows, and defines the status S t = 0, then there is no prohibition on sales; S t = 1, then there is a prohibition on sales, and the value range of t is from 1 to T; define the penalty factor Calculate the number of times the entity has been marked as prohibited from sales during the historical window period; define the reward factor Represents the proportion of the longest length that the entity has been continuously marked as not prohibited from sales during the historical window period, where L represents the historical time window starting from t = 1, S t = 0 is the maximum continuous length; obtain the prohibited sales duration d = (1 - pass)*r*(1 + max{p - b, 0}), where pass is the passing rate of the prohibited sales entity; A sales ban execution module that, based on the intersection of the actual situation and the model prediction situation in the confusion matrix, obtains four situations of true negative, false negative, false positive, and true positive. Among them, negative corresponds to the model not performing a sales ban, positive corresponds to the model performing a sales ban, true corresponds to the model prediction result being consistent with the actual situation, and false corresponds to the model prediction result being inconsistent with the actual situation; define the improvement rate as the proportion of actual interceptions that enter the sales ban, that is, TN / All; define the misban rate as the proportion of those that actually pass but enter the sales ban, that is, FN / All; traverse all sales ban thresholds according to the preset step interval, and calculate the curve of the improvement rate and misban rate under the threshold; calculate the marginal utility of improvement and the marginal cost of misban, and find the sales ban threshold with a large improvement rate and a small misban rate, which is considered the most valuable sales ban threshold; use the sales ban threshold with a large improvement rate and a small misban rate as the finest-grained sales ban threshold, and the sales ban thresholds at the remaining levels decay uniformly at a step of the sales ban threshold with a large improvement rate and a small misban rate / N; The deduplication and merging module obtains unique entities subject to sales bans through a deduplication and merging algorithm.

6. A dynamic adjustment device for air ticket reservation, characterized in that, It includes: A processor; A memory that stores executable instructions for the processor; Wherein, the processor is configured to execute the steps of the dynamic adjustment method for air ticket reservation according to any one of claims 1 to 4 by executing the executable instructions.

7. A computer-readable storage medium for storing a program, characterized in that, When the program is executed by the processor, it implements the steps of the dynamic adjustment method for air ticket reservation according to any one of claims 1 to 4.

Images