Multi-dimensional airfare validity verification method

Abstract

The application relates to the technical field of airfare verification, and discloses a multi-dimensional airfare validity verification method, which comprises the following steps: element disassembling airfare data, determining an airfare knowledge graph by correlating the airfare constituting relationship of each airfare element node, extracting constraint nodes of airfare changes based on the airfare knowledge graph, determining the constraint relationship among the airfare element nodes according to the rule content corresponding to each rule node, determining the correlation relationship among the airfare element nodes based on an association rule algorithm, mapping the element disassembling result to the airfare element nodes corresponding to the airfare knowledge graph, and performing constraint matching on each difference value based on an airfare element constraint data set; the validity state of each airfare element node is determined based on the constraint matching result; and the airfare validity verification result is determined based on the validity state of all the airfare element nodes. The application ensures the reliability of the validity verification of multi-dimensional airfares.

Description

Technical Field

[0001] This invention relates to the field of airfare verification technology, and more specifically, to a multi-dimensional method for verifying the validity of airfares. Background Technology

[0002] As air passenger transport systems evolve towards openness, service orientation, and real-time transaction processing, air fare processing is no longer limited to whether the back-end fare calculation results are correct. Instead, it further requires ensuring that the quoted price is consistent with the fare status at the time of order placement in the actual user transaction process.

[0003] Existing patent CN119722201A discloses a method, device, storage medium, and electronic device for verifying international airfare prices. This method constructs a query request or an input request, sending the same business request to both the host-side system and the open-side system. By comparing the processing results from both sides, the fare calculation function is verified. This type of verification focuses on the consistency of system processing results and is geared towards functional correctness comparisons during testing and system migration. It lacks specific handling for the fare status changes caused by the time difference between checking prices and placing an order in real-world user ticketing scenarios. Furthermore, because airfare data has dynamic fluctuation characteristics, the quotes obtained by users are usually from cached results or previous query results. Fare elements such as flight availability and tax amounts can change, and the lack of real-time verification capabilities at status transition points leads to insufficient verification of airfare validity.

[0004] Therefore, it is necessary to design a multi-dimensional airfare validity verification method to solve the problems existing in the current technology. Summary of the Invention

[0005] In view of this, the present invention proposes a multi-dimensional method for validating airfare validity, which aims to solve the problems of insufficient processing of fare status changes caused by the time difference between checking prices and placing orders in the ticket purchase scenario, and the lack of ability to perform real-time verification at the status transition node, resulting in inadequate verification of airfare validity.

[0006] This invention proposes a multi-dimensional method for verifying the validity of airfares, including: Acquire flight identification data, cabin class data, and fare data, and decompose the fare data into elements to determine the fare impact data. Construct fare element nodes based on the flight identification data, cabin class data, and fare impact data, and associate the fare composition relationships of each fare element node to determine the fare knowledge graph. Based on the freight rate knowledge graph, constraint nodes for freight rate changes are extracted. The constraint nodes include at least fuel surcharge rule nodes, tax calculation rule nodes, and cabin allocation rule nodes. The constraint relationships between each freight rate element node are determined according to the rule content corresponding to each rule node. The association relationships between each freight rate element node are determined based on the association rule algorithm. The freight rate element constraint dataset is determined based on the constraint relationships and association relationships. The system acquires flight data corresponding to the current flight, decomposes the flight data into elements, maps the result of the element decomposition to the fare element nodes corresponding to the fare knowledge graph, determines the target fare knowledge graph, compares each fare element node in the target fare knowledge graph based on the node type to determine the difference value of all nodes, performs constraint matching on each difference value based on the fare element constraint dataset, and determines the validity status of each fare element node based on the constraint matching result. The validity verification result of the fare validity is determined based on the validity status of all fare element nodes. When there are fare element nodes that do not satisfy the constraints, the corresponding fare difference type and the corresponding difference node identifier are output.

[0007] Furthermore, when decomposing the fare data into elements and constructing fare element nodes, the process includes: the fare impact data includes ticket price data, tax data, and fuel surcharge data; parsing the fare data to extract the ticket price field, tax field, and fuel surcharge field; classifying and labeling each field according to its type to determine the ticket price dataset, tax dataset, and fuel surcharge dataset; performing segment matching on the ticket price dataset based on the flight identification data; and associating the ticket price dataset with different cabin classes based on the cabin class data to construct the ticket price node.

[0008] Furthermore, when decomposing the fare data into elements and constructing fare element nodes, the method further includes: splitting the tax dataset into sub-items, classifying taxes based on tax codes, constructing several tax sub-nodes, and determining the correspondence between tax sub-nodes and fare nodes; parsing the fuel surcharge dataset according to rules, determining the correspondence between fuel surcharges and flight segment distances or flight area divisions, constructing fuel surcharge nodes, and combining the fare nodes, tax sub-nodes, and fuel surcharge nodes to construct the fare element nodes.

[0009] Furthermore, in determining the fare knowledge graph, the process includes: identifying relationships among the fare element nodes; determining fare calculation relationship edges based on the summation relationship between the fare node and the tax sub-node; determining additional relationship edges based on the correspondence between the fuel surcharge node and the flight identifier data; determining cabin class constraint relationship edges based on the cabin class matching relationship between the cabin class data and the fare node; and determining the fare knowledge graph based on each fare element node and its corresponding relationship edges.

[0010] Furthermore, in determining the constraint relationships between each fare element node, the process includes: traversing all nodes in the fare knowledge graph, filtering constraint nodes according to node type, parsing the fuel surcharge rule node to extract the corresponding rule parameters between the fuel surcharge and the airspace division or segment distance, decomposing the tax calculation rule node to extract the corresponding charging benchmark for each tax type, extracting constraints from the cabin allocation rule node to determine the matching conditions between the number of cabins and the fare node, and determining the constraint relationships between each fare element node based on the rule parameters corresponding to each rule node.

[0011] Furthermore, in determining the correlation between each fare element node, the process includes: identifying historical fare element nodes, extracting data from the historical fare element nodes to determine a fare element dataset, dividing the fare element dataset into several data items, determining the support of each data item, determining a frequent data item header table based on the support, constructing a corresponding conditional pattern base from bottom to top based on the frequent data item header table to determine a frequent fare element itemset, and determining the correlation between each fare element node based on the frequent fare element itemset.

[0012] Furthermore, in determining the target fare knowledge graph, the process includes: parsing fields, matching flight segments, and splitting taxes in the flight data to determine real-time fare nodes, real-time tax sub-nodes, and real-time fuel surcharge nodes, and mapping each real-time node to the corresponding fare element node in the fare knowledge graph to determine the target fare knowledge graph.

[0013] Furthermore, when comparing and determining the difference values ​​of each fare element node, the process includes: extracting each fare element node and its corresponding real-time fare element node, calculating the difference values ​​for each of the fare node, tax sub-node, and fuel surcharge node, determining the difference values ​​for each node, determining the node difference value dataset based on all node difference values, performing item-by-item matching on the node difference value dataset based on the fare element constraint dataset, determining the constraint matching results corresponding to each difference value, and marking the validity status of the corresponding fare element node based on the matching results.

[0014] Furthermore, when outputting the fare difference type and difference node identifier, the method includes: determining the difference category based on the node type corresponding to the difference value, wherein the difference category includes at least the fare difference category, the tax difference category, and the fuel surcharge difference category, and binding the difference category with the corresponding node type identifier to determine the difference node identifier.

[0015] Furthermore, the fare knowledge graph includes at least a fare layer, a tax layer, a fuel surcharge layer, and a cabin class layer, with each layer's nodes connected based on a relationship identifier, and each node recording the corresponding node type identifier, numerical field, and association relationship identifier.

[0016] Compared with existing technologies, the advantages of this invention are as follows: By constructing a fare knowledge graph, a multi-dimensional structural integration of airfares is achieved, transforming scattered flight identification data, cabin class data, and fare impact data into a relational node system, avoiding the limitations of simply comparing the consistency of system processing results. Furthermore, by extracting fuel surcharge rule nodes, tax calculation rule nodes, and cabin allocation rule nodes to form constraint relationships, and combining these with association rule algorithms to mine these relationships and construct a fare element constraint dataset, a multi-dimensional verification basis combining rule constraints and data association is formed. This avoids the one-sidedness of relying solely on fixed rules, improving the comprehensiveness and reliability of verification. By acquiring current flight data and mapping it to generate a target fare knowledge graph, real-time fare verification is achieved in the user's ticket purchase transaction chain. This solves the problem of changes in fare elements such as cabin class, taxes, and fuel surcharges due to time differences between price inquiry and order placement, ensuring the reliability of real-time verification, achieving quantitative verification of fare element fluctuations, and guaranteeing the accuracy of verification results. Simultaneously, the quoted price remains consistent with the fare status at the time of order placement, improving the targeting and reliability of airfare validity verification. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a multi-dimensional airfare validity verification method provided in an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] See Figure 1 As shown in some embodiments of this application, a multi-dimensional airfare validity verification method includes: S100: Acquire flight identification data, cabin class data, and fare data, decompose the fare data into elements, determine the fare impact data, construct fare element nodes based on flight identification data, cabin class data, and fare impact data, and associate the fare composition relationships of each fare element node to determine the fare knowledge graph; S200: Extract constraint nodes for fare changes based on the fare knowledge graph. The constraint nodes include at least fuel surcharge rule nodes, tax calculation rule nodes, and cabin allocation rule nodes. Determine the constraint relationship between each fare element node based on the rule content corresponding to each rule node, and determine the association relationship between each fare element node based on the association rule algorithm. Determine the fare element constraint dataset based on the constraint relationship and the association relationship. S300: Obtain the flight data corresponding to the current flight, decompose the flight data into elements, map the result of the element decomposition to the corresponding fare element node of the fare knowledge graph, determine the target fare knowledge graph, compare each fare element node in the target fare knowledge graph according to the node type to determine the difference value of all nodes, and perform constraint matching on each difference value based on the fare element constraint dataset, and determine the validity status of each fare element node based on the constraint matching result; S400: Determine the validity verification result of fares based on the validity status of all fare element nodes, and when there are fare element nodes that do not meet the constraints, output the corresponding fare difference type and the corresponding difference node identifier.

[0022] Specifically, flight identification data, cabin class data, and fare data are obtained through existing aviation data platforms such as civil aviation reservation systems and fare management systems. Flight identification data represents a set of specific information used to uniquely identify each airline flight, including flight number, flight segment information, etc. This information allows for precise location of a specific flight, avoiding flight confusion. Cabin class data represents all class options and corresponding configuration information for a particular flight, including cabin class codes (e.g., Economy Class Y, Business Class C, First Class F) and the number of seats available in each class. Fare data represents the overall fare-related data for a particular flight and cabin class, including the base fare, various taxes and fees, fuel surcharges, and all other fare-related information. Fare data is broken down into elements to separate all sub-data that directly affects the final fare amount, i.e., determining fare impact data. Fare impact data represents various sub-data that directly affect the final airfare amount. For example, the fare impact data obtained after decomposition includes three sub-data categories: base fare, airport construction fee, and fuel surcharge. Based on the acquired flight identification data, cabin class data, and decomposed fare impact data, fare element nodes are constructed one by one. A fare element node represents the smallest data unit (i.e., node) in the knowledge graph that can be independently stored and analyzed in relation to various types of fare impact data, flight-related data, and cabin class-related data after decomposition. Each fare element node corresponds to only one independent fare-related element and contains all the information of that element. For example, the fare element node includes "Flight Identifier Node (PEK→PVG, 08:00-10:30)", "Cabin Class Node (Y class, 30 available, full price)", "Base Fare Node (1200 yuan, associated with Y class)", "Airport Construction Fee Node (50 yuan)", and "Fuel Surcharge Node (200 yuan)". By sorting out and linking the fare composition relationships between various fare element nodes, the fare composition relationships represent the specific correspondences such as combination, subordination, and calculation between various fare element nodes to form the complete fare for a certain flight and a certain cabin class. For example, "fare node + airport construction fee node + fuel surcharge node = complete fare". By using these composition relationships as relational edges connecting each node, all fare element nodes are integrated and linked, and finally the fare knowledge graph is determined. This comprehensively integrates the structured knowledge network of flights, cabin classes, and various fare-influencing factors, realizing the unified integration of fare-related factors and improving the comprehensiveness of fare effectiveness verification.

[0023] Specifically, based on the constructed fare knowledge graph, constraint nodes that determine changes in fare amounts and applicable conditions are selected and extracted. These constraint nodes represent nodes in the fare knowledge graph used to limit the numerical range, matching conditions, and calculation rules of fare element nodes. These nodes directly determine the validity of the fare and include at least three types of nodes: fuel surcharge rule nodes, tax calculation rule nodes, and cabin allocation rule nodes. Among them, fuel surcharge rule nodes represent rule-type nodes that record the standards, calculation logic, and scope of application of aviation fuel surcharges. For example, this node stores specific rules such as "When the flight segment distance is ≤ 800 km, the fuel surcharge is 80 yuan; when the flight segment distance is ≤ 1500 km, the fuel surcharge is 200 yuan; when the flight segment distance is > 1500 km, the fuel surcharge is 300 yuan." The tax calculation rule node represents a type of rule node that records the calculation basis, collection standards, and applicable objects of various aviation taxes and fees (such as the Civil Aviation Development Fund and Airport Construction Fee). The cabin allocation rule node represents a type of rule node that records the available quantity of each cabin class for a certain flight, the matching conditions between cabin class and ticket price, and the rules for cabin class status changes. For example, a certain flight has a total of 50 available Y cabin seats. When the available quantity is ≤5, the ticket price increases by 10%. The Y cabin only matches the full-price base ticket price and is not subject to specific rules such as discounted tickets. After extracting the constraint nodes, based on the rule content in each rule node, the constraint relationship between each fare element node is sorted out and determined. That is, the rule restrictions that each fare element node follows. For example, according to the rule of the fuel surcharge rule node, the value of the fuel surcharge node must be equal to 200 yuan. Since the flight segment distance is 1200 kilometers, it falls within the 800-1500 kilometer range. That is, the constraint relationship between the fuel surcharge node and the fuel surcharge rule node. Meanwhile, association rule algorithms are used to determine the relationships between various fare element nodes. For example, by analyzing historical flight data from the past six months, it is found that the three nodes of Y class + base fare of 1200 yuan + fuel surcharge of 200 yuan usually appear together. This relationship can serve as a supplement to the constraint relationship, thereby assisting in judging the validity of the fare. By constructing fare element nodes and forming a fare element constraint dataset, subsequent verification can cover every fare element node, avoiding the problems of single field comparison and incomplete element coverage in existing technologies. Moreover, through node-level construction, the differences between various elements can be accurately calculated, ensuring the reliability of validating airfares.

[0024] Understandably, when a user initiates a ticket purchase order, the system retrieves the flight data for the current flight. This data represents the latest real-time data for the flight at the moment the user places the order, including the number of available seats, the fuel surcharge, and taxes. The system then performs the same element decomposition as before, mapping each of the resulting real-time subdivided data points to the corresponding fare element nodes in the previously constructed fare knowledge graph. This forms the target fare knowledge graph, which incorporates the fare knowledge graph containing the real-time flight data elements at the moment the user places the order. This serves as a reference knowledge graph for real-time verification of fare validity. Based on node type, each fare element node in the target fare knowledge graph is compared one by one with the corresponding node in the fare knowledge graph. The system calculates and determines the difference value for all nodes. This difference value represents the numerical or state deviation between the real-time fare element node in the target fare knowledge graph and the fare element node in the fare knowledge graph. For example, if the original fuel surcharge node value is 200 yuan and the real-time node value is 220 yuan, the corresponding node difference value is 20 yuan. Based on the constructed fare element constraint dataset, constraint matching is performed on the difference value of each node, that is, comparison and verification one by one to determine whether the difference value meets the constraint requirements. For example, for the node difference value of 20 yuan for fuel surcharge, it is compared with the constraint condition in the constraint dataset that the fuel surcharge for Y cabin must be 200 yuan. For the difference value of -5 available cabins, it is compared with the constraint condition in the constraint dataset that the available cabins for Y cabin are ≥0. The validity status of each fare element node is determined according to the constraint matching result, that is, whether it meets the constraint conditions in the fare element constraint dataset. For example, the fuel surcharge node is determined to be "invalid" because the difference value does not meet the constraint, while the cabin node is determined to be "valid" because the difference value meets the constraint. The system aggregates the validity status of all fare element nodes and comprehensively determines the overall fare validity verification result. If all fare element nodes are "valid," the fare is deemed valid. If at least one fare element node is "invalid" (i.e., does not satisfy the constraint relationship), the fare is deemed invalid. Simultaneously, the system outputs the corresponding fare difference type and the corresponding difference node identifier. The fare difference type indicates the category to which the node with the deviation belongs; for example, a fuel surcharge node deviation corresponds to "fuel surcharge difference," and a cabin class node deviation corresponds to "cabin class difference." The difference node identifier is a unique marker used to accurately locate the specific fare element node with the deviation; for example, fuel surcharge node -001. The difference node identifier allows for quick identification of invalid nodes, enabling real-time verification of fare status in the user's ticket purchase transaction process. This ensures the reliability of multi-dimensional airline fare validity verification. Furthermore, by verifying and reporting invalidity in real time, it avoids issues such as order failures due to fare changes.

[0025] In some embodiments of this application, when decomposing fare data into elements and constructing fare element nodes, the process includes: fare impact data including ticket price data, tax data, and fuel surcharge data; parsing the fare data to extract the ticket price field, tax field, and fuel surcharge field; classifying and labeling each field according to its type to determine the ticket price dataset, tax dataset, and fuel surcharge dataset; performing segment matching on the ticket price dataset based on flight identifier data; and associating the ticket price dataset with different cabin classes based on cabin class data to construct the ticket price node.

[0026] In some embodiments of this application, when decomposing fare data into elements and constructing fare element nodes, the method further includes: splitting the tax dataset into sub-items, classifying taxes based on tax codes, constructing several tax sub-nodes, and determining the correspondence between tax sub-nodes and fare nodes; parsing the fuel surcharge dataset according to rules, determining the correspondence between fuel surcharges and flight segment distances or flight area divisions, constructing fuel surcharge nodes, and combining fare nodes, tax sub-nodes, and fuel surcharge nodes to construct fare element nodes.

[0027] Specifically, the fare data represents the basic passenger fare data for corresponding cabin classes on airline flights; the tax data represents the data related to various statutory taxes and fees levied on civil aviation passenger transport; and the fuel surcharge data represents the data related to additional fees collected based on fuel price fluctuations. The fare data is parsed, breaking down the overall fare data into independent fields, extracting the fare field, tax field, and fuel surcharge field. Then, each field is categorized and labeled according to its type, adding classification tags to different types of fields to distinguish the fare component categories to which each field belongs. This results in the determination of independent fare datasets, tax datasets, and fuel surcharge datasets. The fare dataset represents all fare fields... The integrated datasets are as follows: the tax dataset represents the dataset formed by integrating all tax fields, and the fuel surcharge dataset represents the dataset formed by integrating all fuel surcharge fields. Based on the flight identifier data, the fare dataset is matched with flight segments. Flight segment matching associates and binds the fare data with the corresponding flight segments in the flight identifier data. For example, the fare data of the segment from point A to point B is associated with the flight identifier of that segment. Then, based on the cabin class data, the fare dataset is associated with cabin class. Cabin class association means binding the fare data with the cabin class data of the corresponding flight, such as economy class or business class. This completes the construction of the fare node, which represents the basic fare information of the corresponding flight segment and cabin class. The tax and fee dataset is split into individual tax types. The overall tax and fee data in the dataset is then divided into individual tax types. The split taxes and fees are then classified based on the tax type codes. The tax type codes are unique codes used in the civil aviation system to distinguish different types of taxes and fees. For example, the Civil Aviation Development Fund and the Airport Construction Fee correspond to different tax type codes. Several tax and fee sub-nodes are constructed through classification. The tax and fee sub-nodes represent the subdivision nodes of information for a single tax type. At the same time, the corresponding subordinate relationship between each tax and fee sub-node and the ticket price node is determined. For example, the tax and fee sub-node of a certain flight only corresponds to the ticket price node of the same flight segment and the same cabin class. The fuel surcharge dataset is analyzed to clarify the calculation basis and implementation standards of the fuel surcharge, and to determine the correspondence between the fuel surcharge and flight segment distance or air area division. Flight segment distance represents the actual mileage of the flight segment, and air area division represents the different flight areas designated by the civil aviation administration. For example, a fixed amount of fuel surcharge corresponds to a flight segment distance of less than 800 kilometers. Domestic air areas and international air areas correspond to different fuel surcharge standards. Fuel surcharge nodes are constructed through this correspondence. Each fuel surcharge node represents a node containing the fuel surcharge amount and corresponding calculation rule information.The completed fare nodes, tax sub-nodes, and fuel surcharge nodes are combined, that is, integrated according to the composition of fares, to construct a complete fare element node. Through multi-level decomposition, classification, and association of fare data, complex comprehensive fare data can be broken down into refined and specialized node units, avoiding information confusion caused by the mixing of different types of fare data. At the same time, the nodes are refined according to dimensions such as tax type, flight segment, cabin class, and flight area, ensuring the reliability of the fare knowledge graph, thereby improving the verification of fare validity and the adaptability to air ticketing scenarios, and thus ensuring the accuracy of node comparison and validity verification.

[0028] In some embodiments of this application, determining the fare knowledge graph includes: identifying relationships between fare element nodes; determining fare calculation relationship edges based on the summation relationship between ticket price nodes and tax sub-nodes; determining additional relationship edges based on the correspondence between fuel surcharge nodes and flight identifier data; determining cabin class constraint relationship edges based on cabin class matching relationship between cabin class data and ticket price nodes; and determining the fare knowledge graph based on each fare element node and its corresponding relationship edges.

[0029] Specifically, relationship identification analyzes the attributes, data content, and actual business logic of each fare element node to identify the corresponding, constraining, or computational relationships between different fare element nodes. This involves determining the fare calculation relationship edge based on the summation relationship between the ticket price node and the tax sub-nodes. The summation relationship indicates that the basic ticket price corresponding to the ticket price node is added to the amount of a single tax type corresponding to each tax sub-node, forming a numerical correlation of the complete fare (excluding fuel surcharges). For example, if the ticket price node amount for a Y class flight is 1200 yuan, and the tax sub-nodes include the Civil Aviation Development Fund node (50 yuan) and the Airport Construction Fee node (50 yuan), the summation relationship between these two and the ticket price node is 1200 yuan + 50 yuan + 50 yuan = 1300 yuan. The fare calculation relationship edge connects the ticket price node with each tax sub-node, clearly showing the computational relationship between ticket price and various taxes, thus clarifying the composition ratio of ticket price and taxes in the complete fare. The fuel surcharge relationship edges are determined based on the correspondence between fuel surcharge nodes and flight identification data. This correspondence indicates that the amount and calculation basis of a fuel surcharge node correspond one-to-one with specific flight identification data (such as flight number and flight segment information). For example, if the fuel surcharge node amount for a flight (location A → location B, flight segment distance 1200 km) is 200 yuan, this fuel surcharge node only corresponds to the flight identification data of that flight and cannot be adapted to other flights. The fuel surcharge relationship edges connect the fuel surcharge nodes and flight identification data, allowing for quick location of the specific flight corresponding to a fuel surcharge and clarifying the scope of application of the fuel surcharge. Cabin class constraint relationship edges are determined based on the cabin class matching relationship between cabin class data and fare nodes. The cabin class matching relationship represents different... There are matching restrictions between the cabin class data (such as economy class Y, business class C) and the corresponding fare node amount and discount standard. For example, Y class corresponds to the full fare node (1200 yuan), and C class corresponds to the business class fare node (2500 yuan). The fare node of Y class cannot match the cabin class data of C class. Cabin class constraint relationship edges are used to connect cabin class data and fare nodes. Through this relationship edge, a certain fare node can be limited to only be adapted to a specific type of cabin class, avoiding the situation where cabin class and fare do not match. All the constructed fare element nodes are integrated with the three types of relationship edges (fare calculation relationship edge, additional relationship edge, cabin class constraint relationship edge), and each relationship edge is accurately connected to the relevant fare element node, thereby forming a structured network, namely the fare knowledge graph. By accurately identifying the relationships between various fare element nodes and constructing customized relationship edges, the fare knowledge graph can clearly present the correlation logic of each fare element. This clarifies the fare calculation rules (the sum of ticket price and taxes) and defines the scope of application of each element (fuel surcharges correspond to specific flights, and ticket prices correspond to specific cabin classes). This reduces verification errors caused by unclear node relationships and further ensures the accuracy of fare verification.

[0030] In some embodiments of this application, determining the constraint relationships between various fare element nodes includes: traversing all nodes in the fare knowledge graph, filtering constraint nodes according to node type, parsing the fuel surcharge rule nodes, extracting the corresponding rule parameters between fuel surcharge and airspace division or segment distance, decomposing the tax calculation rule nodes, extracting the corresponding charging benchmarks for each tax type, extracting constraints from the cabin allocation rule nodes, determining the matching conditions between cabin quantity and fare nodes, and determining the constraint relationships between various fare element nodes based on the rule parameters corresponding to each rule node.

[0031] Specifically, each node in the graph is reviewed and retrieved sequentially to identify constraint nodes. After identifying constraint nodes, the fuel surcharge rule nodes are parsed to extract the corresponding rule parameters between the fuel surcharge and the airspace division or flight segment distance. These corresponding rule parameters represent the specific quantitative standards for the fuel surcharge amount as the airspace or flight segment distance changes. For example, the fuel surcharge is 80 yuan when the flight segment distance is less than or equal to 800 kilometers and 200 yuan when it is greater than 800 kilometers. The intervals of 80 yuan, 200 yuan, and 800 kilometers are all such rule parameters. The tax calculation rule nodes are then decomposed to extract the corresponding billing benchmarks for each tax type. The billing benchmark represents the basis for calculating the amount of a certain civil aviation tax. For example, the billing benchmark for the Civil Aviation Development Fund and Airport Construction Fee is a fixed charge per passenger, while the billing benchmark for some international flight taxes is a percentage of the base ticket price. Constraints are extracted from the cabin allocation rule nodes to identify restrictive judgment conditions, thereby determining the matching conditions between cabin quantity and fare nodes. This means defining the requirements that must be met for the remaining cabin quantity and the corresponding fare node to be effective and applicable. For example, a flight can only apply the corresponding basic fare node if the remaining economy class cabin quantity is greater than or equal to 15 seats. Based on the rule parameters, billing benchmarks, and matching conditions corresponding to the fuel surcharge rule nodes, tax calculation rule nodes, and cabin allocation rule nodes, the restrictions between each fare element node are identified and determined. This provides a precise basis for subsequent fare element difference matching and validity determination, thus ensuring the reliability of fare verification.

[0032] In some embodiments of this application, determining the association between various fare element nodes includes: determining historical fare element nodes, extracting data from historical fare element nodes to determine a fare element dataset, dividing the fare element dataset into several data items, determining the support of each data item, determining a frequent data item header table based on the support, constructing a corresponding conditional pattern base from bottom to top based on the frequent data item header table to determine the frequent itemset of fare elements, and determining the association between various fare element nodes based on the frequent itemset of fare elements.

[0033] Specifically, the historical fare element nodes are determined. These nodes can be identified based on historical fares, taxes, and fuel surcharges for different flight classes and time periods. Full data extraction is performed on these nodes, and the fare element dataset is determined by retrieving attribute values ​​from each node. The Fp-growth algorithm is then used to divide the dataset into several data items. For example, economy class, a base fare of 1200 yuan, and a fuel surcharge of 200 yuan can be treated as individual data items, or combined as a composite data item. Finally, the support of each data item is determined; support measures the frequency of occurrence of that data item. The frequency quantification index is determined based on the support of all data items to establish a frequent data item header table. This header table records all frequent data items that meet the support requirements, their corresponding node identifiers, and the index header of the support values. Based on this header table, a bottom-up traversal approach is used to construct a corresponding conditional pattern base for each frequent data item in the header table. The conditional pattern base is a set of prefix data paths formed by using frequent data items as suffixes. The frequent itemset of the fare element clarifies the potential correspondence patterns of stable coexistence and mutual accompaniment among the nodes of each fare element, and supplements the association judgment criteria beyond hard constraints, thereby improving the robustness of fare validity verification.

[0034] In some embodiments of this application, determining the target fare knowledge graph includes: parsing fields, matching flight segments, and splitting taxes in flight data to determine real-time fare nodes, real-time tax sub-nodes, and real-time fuel surcharge nodes, and mapping each real-time node to the corresponding fare element node in the fare knowledge graph to determine the target fare knowledge graph.

[0035] Specifically, the first step is to obtain the flight data corresponding to the time the user places an order. The flight data represents the latest business data such as real-time fares, flight segments, cabin classes, and taxes for the flight at the current transaction node. The flight data is then processed sequentially by parsing fields, matching flight segments, and splitting taxes. The processes of parsing fields, matching flight segments, and splitting taxes are consistent with the previous processing steps and will not be repeated here. The real-time fare node represents the node containing the real-time basic fare information for the corresponding cabin class of the flight at the current time. The real-time tax sub-node represents the subdivided node containing the real-time amount information for a single tax type at the current time. The real-time fuel surcharge node represents the node containing the real-time fuel surcharge amount and corresponding rule information for the flight at the current time. Each real-time node is mapped one-to-one to the matching fare element node in the pre-built fare knowledge graph according to node type, flight segment, and cabin class attributes. The node structure and relationship connections of the graph remain unchanged, and only the real-time data content within the node is updated. The target fare knowledge graph retains the structure and relationship logic of the fare knowledge graph. By performing standardized field parsing, flight segment matching, and tax type splitting on the real-time flight data, real-time nodes that perfectly match the node structure of the fare knowledge graph can be generated. At the same time, through node mapping, the loading of real-time data is completed without destroying the original graph relationship logic, laying a standardized comparison basis for the one-to-one comparison of each fare element node and improving the accuracy and reliability of fare validity verification.

[0036] In some embodiments of this application, when comparing and determining the difference value of each fare element node, the process includes: extracting each fare element node and the corresponding real-time fare element node, calculating the difference value of each of the fare node, tax sub-node, and fuel surcharge node, determining the difference value of the corresponding node, determining the node difference value dataset based on the difference values ​​of all nodes, matching the node difference value dataset item by item based on the fare element constraint dataset, determining the constraint matching result corresponding to each difference value, and marking the validity status of the corresponding fare element node according to the matching result.

[0037] Specifically, fare element nodes are extracted from the constructed fare knowledge graph, and corresponding real-time fare element nodes are extracted from the target fare knowledge graph. Real-time fare element nodes represent nodes generated at the moment a user initiates an order request, carrying data related to the current flight's real-time fare. After synchronizing the retrieval of these two types of corresponding nodes, for the three types of nodes with numerical attributes—fare nodes, tax sub-nodes, and fuel surcharge nodes—the differences between the original node values ​​and the real-time node values ​​are calculated one by one. This step-by-step calculation determines the difference value for each type of node, thus reflecting the actual fluctuation range of fare elements between the query and order placement times. Subsequently, the difference values ​​calculated for all fare element nodes are summarized to form a node difference value dataset. This dataset integrates at least all fare node, tax sub-node, and fuel surcharge node difference values, comprehensively presenting the fluctuation of all fare elements. Based on the pre-constructed fare element constraint dataset, the nodes... Each difference value in the point difference value dataset is matched and verified item by item. This means that the constraint rules, allowable fluctuation range, and other conditions corresponding to that node are compared and verified one by one. The constraint matching result indicates whether the difference value meets the requirements of the freight factor constraint dataset. Based on the constraint matching result, the corresponding freight factor node is marked to determine its validity status. The validity status indicates whether the real-time data of the freight factor node meets the constraint requirements. By calculating the difference for each of the three types of freight factor nodes, the dynamic fluctuation of freight rates can be transformed into quantified difference values. At the same time, by summarizing them to form a node difference value dataset, the fluctuation of all freight factor factors can be comprehensively covered. Then, by combining it with the freight factor constraint dataset for item-by-item matching, the difference can be determined according to the established rules, avoiding the errors caused by human subjective judgment. This accurately marks the validity status of each node, providing a clear and reliable basis for judgment on the overall verification results and the location of difference nodes, and improving the accuracy of freight rate validity verification.

[0038] In some embodiments of this application, when outputting the fare difference type and difference node identifier, the method includes: determining the difference category based on the node type corresponding to the difference value, wherein the difference category includes at least the fare difference category, the tax difference category, and the fuel surcharge difference category, and binding the difference category and the corresponding node type identifier to determine the difference node identifier.

[0039] In some embodiments of this application, the fare knowledge graph includes at least a fare layer, a tax layer, a fuel surcharge layer, and a cabin class layer, and the nodes of each layer are connected based on relationship identifiers, and each node records the corresponding node type identifier, numerical field, and association relationship identifier.

[0040] Specifically, abnormal deviations are categorized based on the node type to which the difference value belongs, thereby determining the difference classification. The difference classification is a type of fare deviation divided according to the fare composition dimension, including at least the fare difference category, the tax difference category, and the fuel surcharge difference category. Among them, the fare difference category indicates the difference type corresponding to the fare node where the numerical deviation does not meet the constraint condition; the tax difference category indicates the difference type corresponding to the tax sub-node where the numerical deviation does not meet the constraint condition; and the fuel surcharge difference category indicates the difference type corresponding to the fuel surcharge node where the numerical deviation does not meet the constraint condition. For example, if it is determined by verification that the difference value of a fare node exceeds the constraint range, then the deviation of that node is classified into the fare difference category. The determined difference category is bound to the node type identifier of the node. The node type identifier is a unique code assigned to each independent fare element node to distinguish the type and specific identity of the node. For example, the node type identifier of the fare node is TIC-001, the node type identifier of the civil aviation development fund tax sub-node is TAX-001, and the node type identifier of the fuel surcharge node is FUE-001. The difference node identifier is a positioning mark that represents the difference type information and the node's unique identity information. It can directly point to the specific abnormal node. For example, the combination "fare difference category-TIC-001" is a difference node that can be accurately located. The fare knowledge graph is divided into at least four layers: fare layer, tax layer, fuel surcharge layer, and cabin class layer. The fare layer aggregates fare nodes for each flight segment and cabin class. The tax layer aggregates tax sub-nodes for various taxes. The fuel surcharge layer aggregates fuel surcharge nodes for each flight. The cabin class layer aggregates cabin class data nodes for each flight. Nodes within each layer and between different layers are interconnected through relationship identifiers. These identifiers are unique markers used to label the computational, hierarchical, and matching relationships between nodes, clearly defining the connections and interaction logic between nodes. Each node in the graph synchronously records a node type identifier, a numerical field, and a relationship identifier. The numerical field represents the field used to store quantitative data such as the fare amount, tax amount, fuel surcharge amount, and number of cabins corresponding to the node. The relationship identifier represents the identifier used to record the constraint relationship, calculation relationship, and matching relationship between the node and other nodes, which is used to clearly reflect the association direction and constraint basis of the node. By classifying differences according to the node type and binding the node type identifier to generate difference node identifiers, the specific node and deviation type of fare anomaly can be quickly and accurately located, improving the location of abnormal fares and improving the accuracy and reliability of the entire fare validity verification process.

[0041] In summary, the beneficial effects of this invention are as follows: By constructing a fare knowledge graph, it achieves multi-dimensional structural integration of airfares, transforming scattered flight identification data, cabin class data, and fare impact data into a relational node system, avoiding the limitations of simply comparing the consistency of system processing results. Furthermore, by extracting fuel surcharge rule nodes, tax calculation rule nodes, and cabin allocation rule nodes to form constraint relationships, and combining these with association rule algorithms to mine these relationships and construct a fare element constraint dataset, a multi-dimensional verification basis combining rule constraints and data association is formed. This avoids the one-sidedness of relying solely on fixed rules, improving the comprehensiveness and reliability of verification. It acquires current flight data and maps it to generate a target fare knowledge graph, enabling real-time fare verification in the user's ticket purchase transaction chain. This solves the problem of changes in fare elements such as cabin class, taxes, and fuel surcharges due to time differences between price inquiry and order placement, ensuring the reliability of real-time verification, achieving quantitative verification of fare element fluctuations, and guaranteeing the accuracy of verification results. Simultaneously, the quoted price remains consistent with the fare status at the time of order placement, improving the targeting and reliability of airfare validity verification.

[0042] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program goods according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0043] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A multi-dimensional method for verifying the validity of airfares, characterized in that, include: Acquire flight identification data, cabin class data, and fare data, and decompose the fare data into elements to determine the fare impact data. Construct fare element nodes based on the flight identification data, cabin class data, and fare impact data, and associate the fare composition relationships of each fare element node to determine the fare knowledge graph. Based on the freight rate knowledge graph, constraint nodes for freight rate changes are extracted. The constraint nodes include at least fuel surcharge rule nodes, tax calculation rule nodes, and cabin allocation rule nodes. The constraint relationships between each freight rate element node are determined according to the rule content corresponding to each rule node. The association relationships between each freight rate element node are determined based on the association rule algorithm. The freight rate element constraint dataset is determined based on the constraint relationships and association relationships. The system acquires flight data corresponding to the current flight, decomposes the flight data into elements, maps the result of the element decomposition to the fare element nodes corresponding to the fare knowledge graph, determines the target fare knowledge graph, compares each fare element node in the target fare knowledge graph based on the node type to determine the difference value of all nodes, performs constraint matching on each difference value based on the fare element constraint dataset, and determines the validity status of each fare element node based on the constraint matching result. The validity verification result of the fare validity is determined based on the validity status of all fare element nodes. When there are fare element nodes that do not satisfy the constraints, the corresponding fare difference type and the corresponding difference node identifier are output.

2. The multi-dimensional airfare validity verification method according to claim 1, characterized in that, When decomposing the fare data into elements and constructing fare element nodes, the process includes: the fare impact data includes ticket price data, tax data, and fuel surcharge data; parsing the fare data to extract the ticket price field, tax field, and fuel surcharge field; classifying and labeling each field according to its type to determine the ticket price dataset, tax dataset, and fuel surcharge dataset; performing flight segment matching on the ticket price dataset based on the flight identification data; and associating the ticket price dataset with different cabin classes based on the cabin class data to construct the ticket price node.

3. The multi-dimensional airfare validity verification method according to claim 2, characterized in that, When decomposing the fare data into elements and constructing fare element nodes, the method further includes: splitting the tax dataset into sub-items, classifying taxes based on tax codes, constructing several tax sub-nodes, and determining the correspondence between tax sub-nodes and fare nodes; parsing the fuel surcharge dataset according to rules, determining the correspondence between fuel surcharges and flight segment distances or flight area divisions, constructing fuel surcharge nodes, and combining the fare nodes, tax sub-nodes, and fuel surcharge nodes to construct the fare element nodes.

4. The multi-dimensional airfare validity verification method according to claim 3, characterized in that, The process of determining the fare knowledge graph includes: identifying relationships among the fare element nodes; determining fare calculation relationship edges based on the summation relationship between the fare node and the tax sub-node; determining additional relationship edges based on the correspondence between the fuel surcharge node and the flight identifier data; determining cabin class constraint relationship edges based on the cabin class matching relationship between the cabin class data and the fare node; and determining the fare knowledge graph based on each fare element node and its corresponding relationship edges.

5. The multi-dimensional airfare validity verification method according to claim 4, characterized in that, When determining the constraint relationships between each fare element node, the process includes: traversing all nodes in the fare knowledge graph, filtering constraint nodes according to node type, parsing the fuel surcharge rule node to extract the corresponding rule parameters between the fuel surcharge and the airspace division or segment distance, decomposing the tax calculation rule node to extract the corresponding billing benchmark for each tax type, extracting constraints from the cabin allocation rule node to determine the matching conditions between the number of cabins and the fare node, and determining the constraint relationships between each fare element node based on the rule parameters corresponding to each rule node.

6. The multi-dimensional airfare validity verification method according to claim 5, characterized in that, When determining the relationships between various fare element nodes, the process includes: identifying historical fare element nodes, extracting data from the historical fare element nodes to determine a fare element dataset, dividing the fare element dataset into several data items, determining the support of each data item, determining a frequent data item header table based on the support, constructing a corresponding conditional pattern base from bottom to top based on the frequent data item header table to determine the frequent itemset of fare elements, and determining the relationships between various fare element nodes based on the frequent itemset of fare elements.

7. The multi-dimensional airfare validity verification method according to claim 6, characterized in that, The process of determining the target fare knowledge graph includes: parsing the fields of the flight data, matching flight segments and splitting taxes, determining the real-time fare node, the real-time tax sub-node and the real-time fuel surcharge node, and mapping each real-time node to the corresponding fare element node in the fare knowledge graph to determine the target fare knowledge graph.

8. The multi-dimensional airfare validity verification method according to claim 7, characterized in that, When comparing and determining the difference values ​​of each fare element node, the process includes: extracting each fare element node and its corresponding real-time fare element node, calculating the difference values ​​for each of the fare node, tax sub-node, and fuel surcharge node, determining the difference values ​​for each node, determining the node difference value dataset based on all node difference values, performing item-by-item matching on the node difference value dataset based on the fare element constraint dataset, determining the constraint matching results corresponding to each difference value, and marking the validity status of the corresponding fare element node based on the matching results.

9. The multi-dimensional airfare validity verification method according to claim 1, characterized in that, When outputting fare difference types and difference node identifiers, the process includes: determining the difference category based on the node type corresponding to the difference value, wherein the difference category includes at least the fare difference category, tax difference category, and fuel surcharge difference category, and binding the difference category with the corresponding node type identifier to determine the difference node identifier.

10. The multi-dimensional airfare validity verification method according to claim 1, characterized in that, The fare knowledge graph includes at least a fare layer, a tax layer, a fuel surcharge layer, and a cabin class layer. The nodes in each layer are connected based on relationship identifiers, and each node records the corresponding node type identifier, numerical field, and association relationship identifier.

Images