A civil aviation passenger profiling method integrating discrete choice model and knowledge graph

Abstract

The present invention proposes a civil aviation passenger portrait method that integrates discrete choice models and knowledge graphs, including: determining the entities and relationships in the field of civil aviation passenger portraits, and constructing the knowledge ontology of the civil aviation passenger portrait graph; designing a virtual waiting experimental system to collect information while simulating the geographical environment and waiting situation of a virtual terminal; calibrating and analyzing the parameters of civil aviation passenger needs based on a discrete choice model, and forming a preliminary selection probability annotation; processing information, and constructing a visualized civil aviation passenger portrait knowledge graph; using a query language to query and analyze the civil aviation passenger portrait knowledge graph, and characterizing and analyzing the individual and group portraits of civil aviation passengers based on knowledge graph technology. This method has a strong explanatory power for civil aviation passenger behavior decisions, while linking the airport environment and individual passengers, accurately perceiving passenger needs, and comprehensively characterizing individual and group portraits of civil aviation passengers.

Description

Technical Field

[0001] The present invention relates to a civil aviation passenger profiling method, in particular to a civil aviation passenger profiling method integrating a discrete selection model and a knowledge graph. Background Art

[0002] With the rapid development of the economy, people's service demands for civil aviation airports have become diversified. In the past, passengers' demand for airport services was limited to simple boarding and waiting procedures, but now terminal passengers' demands also include non-aviation services such as shopping, dining, and entertainment. How to achieve multi-dimensional perception of passenger needs is a common difficulty in the current civil aviation industry. The key to multi-dimensional perception of demand lies in the comprehensive extraction of passengers' heterogeneous demand characteristics. Passenger portraits, as a common tool for portraying passenger characteristics, can abstract passenger information into various labels, and use labels to concretize the passenger image, thereby providing targeted services.

[0003] Existing passenger portrait research mainly uses discrete choice models to infer passenger behavior probabilities and characterizes the civil aviation passenger group portrait based on the inference results. This model is widely used and has strong explanatory power in passenger behavior decision-making. However, it is based on the maximum utility theory and assumes that passengers' choices are completely rational. It ignores the heterogeneity of individual civil aviation passengers and makes it difficult to focus on the personalized needs of each passenger. It is also difficult to support the storage requirements of large-scale passenger demand data.

[0004] Based on the above problems, it is urgent to find a new type of method that can perceive the multi-dimensional needs of passengers in a changing airport environment while integrating the strong interpretability of discrete choice models. While ensuring strong storage capabilities for large-scale data, it can link the environment and passengers to further perceive passenger portraits. Summary of the invention

[0005] Purpose of the invention: The technical problem to be solved by the present invention is to provide a civil aviation passenger profiling method that integrates discrete choice models and knowledge graphs in response to the deficiencies in the prior art.

[0006] In order to solve the above technical problems, the present invention discloses a civil aviation passenger profiling method integrating a discrete selection model and a knowledge graph, and the specific steps include:

[0007] Step 1: Determine the entities and relationships in the field of civil aviation passenger portraits, and construct the knowledge ontology of civil aviation passenger portraits in a top-down manner;

[0008] Step 2: Design a virtual terminal waiting experimental system to collect basic passenger attribute information, historical travel information, travel preference information, and activity trajectory information in the airport while simulating the virtual terminal geographical environment and waiting situation;

[0009] Step 3: calibrate and analyze the parameters of civil aviation passenger demand based on the discrete choice model, and form a preliminary selection probability label;

[0010] Step 4: Process the basic attribute information, historical travel information, travel preference information and activity trajectory information in the airport collected in step 2, and construct a visual knowledge graph of civil aviation passenger portraits;

[0011] Step 5: Use query language to query and analyze the civil aviation passenger portrait knowledge graph constructed in step 4, and characterize and analyze the individual and group portraits of civil aviation passengers based on knowledge graph technology.

[0012] Furthermore, the knowledge ontology of the civil aviation passenger portrait map is constructed in a top-down manner as described in step 1, specifically including:

[0013] Design a first-level ontology framework, including basic passenger attributes, context, and airport environment;

[0014] Design the secondary ontology, including:

[0015] The secondary ontology of basic passenger attributes includes: socio-demographic characteristics, travel characteristics, travel preference characteristics and activity chain information. The secondary ontology of basic passenger attributes is further subdivided into attribute entities, including at least: age, gender, occupation, income, position, dietary preferences and shopping preferences;

[0016] The secondary ontology of the airport environment includes: grids and activity points. Based on the secondary ontology of the airport environment, attribute entities are subdivided downwards, including at least: grid coordinates, grid spacing, activity point attributes, store brands, and store types;

[0017] The secondary ontology of the context includes: attributes of accompanying persons, flight type, flight departure time, travel purpose, waiting time, and number of checked baggage.

[0018] Furthermore, the design of a terminal virtual waiting experimental system described in step 2 specifically includes:

[0019] Design an intention survey questionnaire to record passengers' basic attribute information, historical travel information, and historical preference information;

[0020] Design a virtual terminal passenger waiting experiment, including virtual airport environment design and random situation experiment design; virtual waiting experiment based on behavioral research, randomly assign virtual situation information and airport environment to passengers, and collect passengers' activity trajectory information under the changing airport environment.

[0021] Furthermore, the parameter calibration and specific analysis of civil aviation passenger demand based on the discrete choice model described in step 3 are performed, and a preliminary selection probability annotation is formed, which specifically includes:

[0022] Assuming that the set of available store types in the airport faced by passenger n is J(i∈J), the utility function of passenger n choosing store type i is expressed as:

[0023]

[0024] Among them, U in represents the utility of passenger n choosing airport store type i; V in represents the observable utility of passenger n choosing store type i; ε in represents the random utility of traveler n choosing store type i;

[0025] The probability P that traveler n chooses store type i in It is expressed as:

[0026] P in =prob(U in >U jn ;i≠j,j∈J)

[0027] Among them, prob represents the probability function of passenger behavior selection, U in represents the utility of passenger n choosing airport store type i, U jn represents the utility of passenger n choosing airport store type j, where j represents the store type and:

[0028] 0≤P in ≤1

[0029] That is, we get:

[0030]

[0031] Among them, Φ() represents the cumulative distribution function of the standard normal distribution, σ represents the variance value, and ε jn represents the random utility of passenger n choosing store type j, V jn represents the observable utility of passenger n choosing store type j, and the probability P in Passengers are labeled as selection probabilities.

[0032] Furthermore, the construction of a visualized civil aviation passenger portrait knowledge graph described in step 4 specifically includes:

[0033] Step 4-1, segmentally labeling the passenger's activity trajectory information in the airport collected in step 2;

[0034] Step 4-2, construct the basic attribute information of passengers, historical travel information, travel preference information and activity trajectory information in the airport collected in step 2 into a triple form of <entity, relationship, entity> or <entity, attribute, attribute value>;

[0035] Step 4-3, import the above triplet data set into storage, and use visualization tools to build a visual civil aviation passenger portrait knowledge graph.

[0036] Furthermore, the description and analysis of individual and group portraits of civil aviation passengers based on knowledge graph technology described in step 5 specifically includes:

[0037] Step 5-1, create a node in the knowledge graph for each passenger, including the passenger's basic attribute characteristics, historical travel characteristics, travel preference characteristics, and activity trajectory characteristics in the airport;

[0038] Step 5-2, define the relationship types between entities and specify attributes for each relationship;

[0039] Step 5-3: Analyze the common characteristics and behavior trends of passenger groups through data mining technology, and identify the key influencing factors among passenger groups;

[0040] Step 5-4: Use advanced query languages ​​to perform data query and pattern recognition, and analyze the correlation between passenger behaviors and preferences, as well as their impact on travel decisions;

[0041] Step 5-5, use graph visualization tools to extract portrait labels of individual and group passengers.

[0042] Furthermore, the step 5-5 uses a graphic visualization tool to extract the portrait labels of individual and group passengers, specifically including:

[0043] Extract passenger individual portrait labels from the same passenger's choice behavior in different airport environments, and perform visual analysis on the passenger individual portrait;

[0044] Extract passenger individual portrait labels from the selection behaviors of passengers with similar attributes in the same airport environment, and perform visual analysis on passenger individual portraits;

[0045] Analyze the distribution of group passengers from the perspective of different types of stores, extract group portrait labels of civil aviation passengers, and visualize and analyze passenger group portraits;

[0046] Analyze the distribution trend of group passengers from the perspective of passenger preferences, extract group portrait labels of civil aviation passengers, and visualize and analyze passenger group portraits.

[0047] Furthermore, the collection of basic attribute information, historical travel information, travel preference information and activity trajectory information of passengers in step 2 specifically includes:

[0048] The basic attribute information of passengers shall be collected, including at least: gender, age, annual disposable income, unit type, position and education level;

[0049] The passenger's historical travel data collection shall include at least: passenger type, number of civil aviation trips, travel cost ratio, ticket class and whether special civil aviation services are required;

[0050] The collection of passengers' travel preference data includes at least: shopping preferences, dining preferences, beverage preferences and office area preferences; the above preferences are different store types;

[0051] The passenger's activity trajectory data collection in the airport includes at least: activity points, activity routes and activity duration;

[0052] Furthermore, in step 2,

[0053] The design of the intention survey questionnaire specifically includes: collecting basic attribute information of passengers, collecting historical travel data of passengers, and collecting historical preference data of passengers;

[0054] The virtual airport environment design specifically includes:

[0055] Design the check-in counter in the virtual airport environment according to the check-in counter layout rules, including: determining the area occupied by the check-in island, including the size of the check-in island itself and the width of the surrounding queuing area; determining the layout location of the check-in island, that is, determining the width of the circulation area before and after the check-in counter; determining the layout method of the check-in island, including the layout method of the check-in island can be divided into centralized layout and decentralized layout;

[0056] Designing a self-service check-in in a virtual environment according to the layout rules of the self-service check-in specifically includes: determining the floor space of the self-service check-in, including the size of the self-service check-in itself and the width of the surrounding queue area; determining the layout location of the self-service check-in, that is, determining the width of the circulation area before and after the self-service check-in counter; determining the layout method of the self-service check-in, including the layout methods of the self-service check-in, which can be divided into centralized layout and decentralized layout; designing the location and layout of the security checkpoint according to the centralized layout / decentralized layout according to the layout rules of the security checkpoint; designing the layout of the non-functional area according to the centralized layout / decentralized layout according to the commercial format ratio of large hub airports at home and abroad and the number of existing large airport stores;

[0057] The base map rules of the virtual terminal map are designed based on the grid labeling method, and all the unit grids of the map are serialized by assigning a sequence value to each unit grid of the map, specifically including:

[0058] Assume that the horizontal length x of the global two-dimensional environment max , the length in the horizontal and vertical directions is y max , the side length of each unit grid is a. If it is less than a unit grid, it is processed as a unit grid. Thus, the two-dimensional geographical environment is divided into a series of unit grids to obtain a two-dimensional grid environment model. The total number of unit grids can be expressed as:

[0059]

[0060] Among them, n x is the number of unit grid columns, n y is the number of unit grid rows, ceil is the rounding operation;

[0061] Assign a value to each unit grid. If the i-th unit grid is an obstacle grid, assign a value to the obstacle grid n. i Assign a value of 1; if the jth unit grid is a free grid, then for the free grid n j Assign a value of 0;

[0062] A rectangular coordinate system is created on the two-dimensional grid environment model, with the upper left corner of the two-dimensional grid environment model as the coordinate origin o, the horizontal direction of the environment model as the x-axis, and the vertical direction of the environment model as the y-axis, and the unit lengths of the two coordinate axes are all integer multiples of the side lengths of the unit grid. Assume that the sequence set is: C = {1, 2, 3, ..., u, u∈N}, if m represents the sequence number, m∈C, and then assign the sequence value m to each grid from top to bottom and from left to right; in the rectangular coordinate system, the rectangular coordinate method and the serial number method correspond to each other. The coordinate position (x m ,y m ), are determined by the sequence number m, as follows:

[0063] x m =(m-1)modn x +1

[0064]

[0065] Among them, mod is the remainder operation, int is the floor operation, and x m The horizontal coordinate when m is y m is the ordinate when the serial number is m;

[0066] After calibrating the grids on the map, set the occupied grid number of each facility according to the layout rules; edit the grid attributes, including at least the occupied time, color, mutation and whether it is clickable;

[0067] Use drawing tools to draw road networks for maps;

[0068] The randomized scenario experimental design specifically includes:

[0069] Select waiting situation factors, including at least: flight type, travel purpose, waiting time, number of checked baggage, accompanying persons and flight schedule, and calibrate corresponding levels for them.

[0070] Furthermore, the segmented annotation of the collected passenger's activity trajectory information in the airport in step 4-1 specifically includes:

[0071] Step 4-1-1, divide the passenger activity trajectory into segments in chronological order, each segment represents a different activity stage of the airport passenger; divide the passenger activity trajectory into segments according to the activity points, and mark the activity sequence, activity distance and activity type of each segment;

[0072] Step 4-1-2, the collected basic attribute information, historical travel information, travel preference information and activity trajectory information in the airport of the passenger are processed into a triple form, including:

[0073] According to the selection probability marked in step 3, as the probability attribute of the passenger's selection, it is processed into a triple form of <passenger, selection probability, probability value>;

[0074] The distance, type and item price of the activity point in the airport environment are taken as activity point attributes and processed into a triple form of <activity point, attribute, attribute value>;

[0075] The relationship between the link activity point and the store type is processed into the triple form of <activity point, belongs to, store type>; the passenger's activity trajectory is segmented and processed into the triple form of <passenger, choice, activity point>; the passenger's basic attributes and travel preference information are marked as the passenger's personal attributes, which are processed into the triple form of <passenger, attribute type, attribute level>.

[0076] Beneficial effects:

[0077] The present invention proposes a civil aviation passenger portrait characterization method based on knowledge graph. While storing large-scale civil aviation passenger portrait data, it integrates discrete choice models and knowledge graph technology. While having strong explanatory power for civil aviation passenger behavioral decisions, it can link the airport environment and individual passengers, accurately perceive the needs of civil aviation passengers, and comprehensively characterize the individual and group portraits of civil aviation passengers.

[0078] Compared with traditional civil aviation passenger profiling methods, this method not only has strong interpretability, but also has rich entities and relationships, such as passengers, airports, flights, time, etc., as well as the complex relationships between them. It helps to observe civil aviation passenger demand profiles from multiple dimensions, identify more information, and more accurately and comprehensively portray passenger characteristics.

[0079] This method can store and process large amounts of data, including structured data, semi-structured data, and unstructured data, and can more comprehensively collect and process multi-dimensional information on civil aviation passenger portraits.

[0080] This method supports query and analysis using query language, as well as visual analysis and exploration. It can query and analyze civil aviation passenger portraits flexibly and conveniently while intuitively displaying civil aviation passenger portrait information. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments, and the above and / or other advantages of the present invention will become more clear.

[0082] Figure 1 This is a flow chart of a method for characterizing civil aviation passenger portraits based on knowledge graph proposed in the present invention.

[0083] Figure 2 This is a schematic diagram of constructing a passenger activity chain proposed by the present invention.

[0084] Figure 3 This is a schematic diagram of a virtual terminal environment proposed by the present invention.

[0085] Figure 4 This is a schematic diagram of a passenger group portrait proposed by the present invention. DETAILED DESCRIPTION

[0086] The present invention provides a method for characterizing a civil aviation passenger portrait, which can simultaneously collect and identify multi-dimensional information related to civil aviation passenger needs, such as passenger attributes, airport environment, and passenger trajectory, and realize multi-dimensional characterization of civil aviation passenger portraits, including the following steps:

[0087] Step 1: Determine the entities and relationships in the field of civil aviation passenger portraits, and construct the knowledge ontology of civil aviation passenger portraits in a top-down manner;

[0088] Step 2: Design a virtual terminal waiting experimental system to collect basic passenger attribute information, historical travel information, travel preference information, and activity trajectory information in the airport while simulating the virtual terminal geographical environment and waiting situation;

[0089] Step 3: calibrate and analyze the parameters of civil aviation passenger demand based on the discrete choice model, and form a preliminary selection probability label;

[0090] Step 4: Process the collected basic passenger attribute information, historical travel information, travel preference information, and activity trajectory information in the airport, and construct a visual civil aviation passenger portrait knowledge graph;

[0091] Step 5: Use query language to query and analyze the constructed civil aviation passenger portrait knowledge graph, and characterize and analyze the individual and group portraits of civil aviation passengers based on knowledge graph technology.

[0092] In this solution, the knowledge ontology of civil aviation passenger portrait map is constructed in a top-down manner, specifically including:

[0093] Design the main framework of the first-level ontology, including basic passenger attributes, context, and airport environment;

[0094] Design the secondary ontology:

[0095] The secondary ontology of basic passenger attributes includes socio-demographic characteristics, travel characteristics, travel preference characteristics, and activity chain information. Based on the secondary ontology, it is further subdivided into attribute entities such as age, gender, occupation, income, position, dietary preferences, and shopping preferences.

[0096] The secondary ontology of the airport environment includes grids and activity points, which are further subdivided into grid coordinates, grid spacing, activity point attributes, store brands, and store types;

[0097] The secondary ontology of the situation includes attribute entities such as accompanying person attributes, flight type, flight departure time, travel purpose, waiting time and number of checked baggage.

[0098] In this scheme, the design of a terminal virtual waiting experimental system specifically includes:

[0099] Design an intention survey questionnaire to investigate passengers' basic attribute information, historical travel information, and historical preference information;

[0100] Design a virtual terminal passenger waiting experiment, including virtual airport environment design and randomized scenario experiment design;

[0101] In the virtual waiting experiment based on behavioral research, passengers are randomly assigned virtual situational information and airport environment, and the activity trajectory information of passengers in the changing airport environment is collected.

[0102] In this scheme, the design of the intention survey questionnaire specifically includes: collecting basic attribute information of passengers, collecting historical travel data of passengers, and collecting historical preference data of passengers;

[0103] The virtual airport environment design specifically includes:

[0104] Design the check-in counters in the virtual airport environment according to the check-in counter layout rules, including: determining the area occupied by the check-in island, including the size of the check-in island itself and the width of the surrounding queuing area;

[0105] Determine the layout of the check-in island, that is, determine the width of the circulation area before and after the check-in counter;

[0106] Determine the layout of the check-in island, including whether the layout of the check-in island can be centralized or decentralized;

[0107] Designing self-service check-in in a virtual environment according to the self-service check-in layout rules includes:

[0108] Determine the floor space of the self-service check-in machine, including the size of the self-service check-in machine itself and the width of the surrounding queue area;

[0109] Determine the location of the self-service check-in machine, that is, determine the width of the circulation area before and after the self-service check-in counter;

[0110] Determine the layout of the self-service check-in machines, including the layout of the self-service check-in machines can be divided into centralized layout and decentralized layout;

[0111] Design the location and layout of security checkpoints according to the centralized / dispersed layout rules;

[0112] According to the commercial format ratio of large hub airports at home and abroad and the number of shops in existing large airports, the layout of non-functional areas is designed according to centralized layout / dispersed layout;

[0113] The base map rules of the virtual terminal map are designed based on the grid labeling method, and all the unit grids of the map are serialized by assigning a sequence value to each unit grid of the map, specifically including:

[0114] Assume that the horizontal length x of the global two-dimensional environment max , the length in the horizontal and vertical directions is y max , the side length of each unit grid is a. If it is less than a unit grid, it is processed as a unit grid. Thus, the two-dimensional geographical environment can be divided into a series of unit grids. The total number of unit grids can be expressed as:

[0115]

[0116] Among them, n x is the number of unit grid columns, n y is the number of unit grid rows, ceil is the rounding operation;

[0117] Assign a value to each unit grid. If the i-th unit grid is an obstacle grid, assign a value to the obstacle grid n. i Assign a value of 1; if the jth unit grid is a free grid, then for the free grid n j Assign a value of 0;

[0118] A rectangular coordinate system is created on the two-dimensional grid environment model, with the upper left corner of the two-dimensional grid environment model as the coordinate origin o, the horizontal direction of the environment model as the x-axis, and the vertical direction of the environment model as the y-axis, and the unit lengths of the two coordinate axes are all integer multiples of the side lengths of the unit grid. Assume that the sequence set is: C = {1, 2, 3, ..., u, u∈N}, if m represents the sequence number, m∈C, then assign the sequence value m to each grid from top to bottom and from left to right;

[0119] In the rectangular coordinate system, the rectangular coordinate method and the serial number method correspond to each other. The coordinate position (x m ,y m ), are determined by the sequence number m, as follows:

[0120] x m =(m-1)modn x +1

[0121]

[0122] Among them, mod is the remainder operation, int is the floor operation, and x m The horizontal coordinate when m is y m is the ordinate when the sequence number is m;

[0123] After calibrating the grid on the map, set the occupied grid number of each facility according to the layout rules;

[0124] Edit the grid properties, including at least the occupied time, color, mutation and clickability;

[0125] Draw the road network for the map using Adobe Illustrator CS6.

[0126] In this scheme, the randomized scenario experimental design includes:

[0127] Select waiting situation factors, including flight type, travel purpose, waiting time, number of checked baggage, accompanying persons, and flight time, and calibrate corresponding levels for them;

[0128] A virtual scenario experiment was designed to track and monitor passengers' intention choices in a virtual terminal environment under the premise of given random passenger scenarios, and capture them as a complete passenger activity trajectory.

[0129] Collection of basic information about passengers, including gender, age, annual disposable income, unit type, position, and education level;

[0130] Passengers’ historical travel data collection, including passenger type, number of civil aviation trips, percentage of public travel expenses, ticket class, and whether special civil aviation services are required;

[0131] Collection of passengers’ historical preference data, including shopping preferences, dining preferences, beverage preferences, and office area preferences;

[0132] Among them, the types of stores preferred for shopping include luxury stores, cosmetics stores, clothing stores, various duty-free stores, specialty stores, home furnishing stores, convenience stores, and leisure parks;

[0133] The types of restaurants preferred by consumers include Chinese fast food, Western fast food, Chinese formal meals, and Western formal meals;

[0134] The types of stores preferred for beverages include milk tea shops, coffee shops, and convenience stores;

[0135] The types of shops preferred by office areas include cafes, restaurants, and rest areas;

[0136] Passengers’ activity trajectory data collection, including activity points, activity routes, and activity duration.

[0137] In this solution, the collection of the changed airport environment data specifically includes:

[0138] Collection of variable scenario data, including flight type, travel purpose, waiting time, number of checked baggage, accompanying persons, and flight schedule;

[0139] Airport environment data collection, including activity point coordinates, activity point attributes, store brands, and store types.

[0140] In this scheme, the discrete choice model is used to calibrate and analyze the civil aviation passenger demand parameters, and form a preliminary selection probability annotation, which specifically includes:

[0141] Assuming that the set of available store types in the airport faced by passenger n is J(i∈J), the utility function of passenger n choosing store type i is expressed as:

[0142]

[0143] Among them, U in represents the utility of passenger n choosing airport store type i; V in represents the observable utility of passenger n choosing store type i; ε in represents the random utility of traveler n choosing store type i;

[0144] The probability P that traveler n chooses store type i in It is expressed as:

[0145] P in =prob(U in >U jn ;i≠j,j∈J)

[0146] Among them, prob represents the probability function of passenger behavior selection, U in represents the utility of passenger n choosing airport store type i, U jn表 represents the utility of passenger n choosing airport store type j, where j represents the store type and:

[0147] 0≤P in ≤1

[0148] That is, we get:

[0149]

[0150] Among them, Φ() represents the cumulative distribution function of the standard normal distribution, σ represents the variance value, and ε jn represents the random utility of passenger n choosing store type j, V jn represents the observable utility of passenger n choosing store type j, and the probability P in Passengers are labeled as selection probabilities.

[0151] In this solution, the construction of a visualized civil aviation passenger portrait knowledge graph specifically includes:

[0152] Segmentally label the collected activity trajectories of civil aviation passengers in the airport environment;

[0153] Process the collected basic passenger attribute information, historical travel information, travel preference information and activity trajectory information in the airport into a triple form of <entity, relationship, entity> or <entity, attribute, attribute value>;

[0154] Import the triple data set into Neo4j to achieve large-scale data storage and build a visual knowledge graph of civil aviation passenger portraits;

[0155] The activity trajectories of civil aviation passengers are marked in sections, including:

[0156] Divide the passenger activity trajectory into segments in chronological order, with each segment representing a different activity stage of airport passengers, such as security check, waiting, and boarding;

[0157] Divide the passenger activity trajectory into segments according to activity points, and mark the activity sequence, activity distance, and activity type of each segment;

[0158] The collected basic passenger attribute information, historical travel information, travel preference information and activity trajectory information in the airport are processed into triples, including:

[0159] The prediction result of the calibrated discrete choice model is the probability attribute of the passenger's choice, which is processed into a triple form of <passenger, probability of choosing a certain store type, probability value>;

[0160] The distance, type, and item price of the activity points in the airport environment are calibrated as the attribute features of the activity points, and processed into a triple form of <a certain activity point, attribute, attribute value>;

[0161] The relationship between the activity point and the store type is processed as a triple of <activity point, belongs to, a certain store type>;

[0162] The passenger's activity trajectory is segmented into a triple form of <passenger, choice, activity point>;

[0163] The basic attributes and travel preference information of passengers are marked as personal attributes of passengers and processed into a triple form of <passenger, attribute type, attribute level>;

[0164] In this solution, the individual and group portraits of civil aviation passengers are described and analyzed based on the knowledge graph technology, specifically including:

[0165] Create a node for each passenger, including their basic attribute characteristics, historical travel characteristics, travel preference characteristics, and activity trajectory characteristics within the airport;

[0166] Define the relationship types between entities, such as <travelers, choices, stores>, and specify attributes for each relationship;

[0167] Analyze the common characteristics and behavioral trends of passenger groups through data mining technology, and identify key influencing factors among passenger groups, such as travel purpose, preference type, etc.;

[0168] Leverage Neo4J’s advanced query languages ​​(such as Cypher) to perform complex data queries and pattern recognition, and analyze the correlation between traveler behaviors and preferences, and their impact on travel decisions;

[0169] Use graph visualization tools to extract portrait labels of individual and group passengers and extract valuable insights to support airlines' service improvements, marketing strategies, and operational optimization.

[0170] In this solution, the use of a graphic visualization tool to extract the portrait labels of individual and group passengers specifically includes:

[0171] Extract passenger individual portrait labels from the same passenger's choice behavior in different airport environments, and perform visual analysis on the passenger individual portrait;

[0172] Extract passenger individual portrait labels from the selection behaviors of passengers with similar attributes in the same airport environment, and perform visual analysis on passenger individual portraits;

[0173] Analyze the distribution of group passengers from the perspective of different types of stores, extract group portrait labels of civil aviation passengers, and visualize and analyze passenger group portraits;

[0174] Analyze the distribution trend of group passengers from the perspective of passenger preferences, extract group portrait labels of civil aviation passengers, and visualize and analyze passenger group portraits.

[0175] Example:

[0176] like Figure 1 As shown, the present application discloses a method for characterizing civil aviation passenger portraits by integrating a discrete selection model and knowledge graph technology, comprising the following steps:

[0177] S110, determine the entities and relationships in the field of civil aviation passenger portraits, and construct the knowledge ontology of civil aviation passenger portraits in a top-down manner;

[0178] S120, design a terminal virtual waiting experimental system, which collects basic passenger attribute information, historical travel information, travel preference information and activity trajectory information in the airport while simulating the virtual terminal geographical environment and waiting situation;

[0179] S130, calibrating and analyzing the parameters of civil aviation passenger demand based on the discrete choice model, and forming a preliminary selection probability labeling;

[0180] S140, processing the collected basic attribute information, historical travel information, travel preference information and activity trajectory information in the airport, and constructing a visualized civil aviation passenger portrait knowledge graph;

[0181] S150, use query language to query and analyze the constructed civil aviation passenger portrait knowledge graph, and characterize and analyze the individual portrait and group portrait of civil aviation passengers based on knowledge graph technology.

[0182] According to an embodiment of the present invention, the knowledge ontology of constructing a civil aviation passenger portrait map in a top-down manner specifically includes:

[0183] Design the main framework of the first-level ontology, including basic passenger attributes, context, and airport environment;

[0184] Design the secondary ontology, including:

[0185] The secondary ontology of basic passenger attributes includes socio-demographic characteristics, travel characteristics, travel preference characteristics, and activity chain information, e.g. Figure 2 As shown in the figure, based on the secondary ontology, it is subdivided into attribute entities such as age, gender, occupation, income, position, dietary preferences, shopping preferences, etc.

[0186] The secondary ontology of the airport environment includes grids and activity points, which are further subdivided into grid coordinates, grid spacing, activity point attributes, store brands, and store types;

[0187] According to an embodiment of the present invention, the design of a terminal virtual waiting experimental system specifically includes:

[0188] Design intention questionnaires to investigate passengers' basic attribute information, historical travel information, and historical preference information;

[0189] Design a virtual terminal passenger waiting experiment, including virtual airport environment design and randomized scenario experiment design;

[0190] In the virtual waiting experiment based on behavioral research, virtual situational information and airport environment are randomly assigned to passengers, and the activity trajectory information of passengers in the changing airport environment is collected.

[0191] According to an embodiment of the present invention, the design of the intention survey questionnaire specifically includes:

[0192] Collection of passengers’ basic attribute information, historical travel data and historical preference data;

[0193] According to an embodiment of the present invention, the virtual airport environment design specifically includes:

[0194] Design the check-in counters in the virtual airport environment according to the check-in counter layout rules, including: determining the area occupied by the check-in island, including the size of the check-in island itself and the width of the surrounding queuing area;

[0195] Determine the layout of the check-in island, that is, determine the width of the circulation area before and after the check-in counter;

[0196] Determine the layout of the check-in island, including whether the layout of the check-in island can be centralized or decentralized;

[0197] Designing self-service check-in in a virtual environment according to the self-service check-in layout rules includes:

[0198] Determine the floor space of the self-service check-in machine, including the size of the self-service check-in machine itself and the width of the surrounding queue area;

[0199] Determine the location of the self-service check-in machine, that is, determine the width of the circulation area before and after the self-service check-in counter;

[0200] Determine the layout of the self-service check-in machines, including the layout of the self-service check-in machines can be divided into centralized layout and decentralized layout;

[0201] Design the location and layout of security checkpoints according to the centralized / dispersed layout rules;

[0202] According to the commercial format ratio of large hub airports at home and abroad and the number of shops in existing large airports, the layout of non-functional areas is designed according to centralized layout / dispersed layout;

[0203] The base map rules of the virtual terminal map are designed based on the grid labeling method, and all the unit grids of the map are serialized by assigning a sequence value to each unit grid of the map, specifically including:

[0204] Assume that the horizontal length x of the global two-dimensional environment max , the length in the horizontal and vertical directions is y max , the side length of each unit grid is a. If it is less than a unit grid, it is processed as a unit grid. Thus, the two-dimensional geographical environment can be divided into a series of unit grids. The total number of unit grids can be expressed as:

[0205]

[0206] Among them, n x is the number of unit grid columns, n y is the number of unit grid rows, ceil is the rounding operation;

[0207] Assign a value to each unit grid. If the i-th unit grid is an obstacle grid, assign a value to the obstacle grid n. i Assign a value of 1; if the jth unit grid is a free grid, then for the free grid n j Assign a value of 0;

[0208] A rectangular coordinate system is created on the two-dimensional grid environment model, with the upper left corner of the two-dimensional grid environment model as the coordinate origin o, the horizontal direction of the environment model as the x-axis, and the vertical direction of the environment model as the y-axis, and the unit lengths of the two coordinate axes are all integer multiples of the side lengths of the unit grid. Assume that the sequence set is: C = {1, 2, 3, ..., u, u∈N}, if m represents the sequence number, m∈C, then assign the sequence value m to each grid from top to bottom and from left to right;

[0209] In the rectangular coordinate system, the rectangular coordinate method and the serial number method correspond to each other. The coordinate position (x m ,y m ), are determined by the sequence number m, as follows:

[0210] x m =(m-1)modn x +1

[0211]

[0212] Among them, mod is the remainder operation, int is the floor operation, and x m The horizontal coordinate when m is ym is the ordinate when the serial number is m;

[0213] After calibrating the grid on the map, set the occupied grid number of each facility according to the layout rules;

[0214] Edit the grid properties, including at least the occupied time, color, mutation and clickability;

[0215] Use drawing tools to draw road networks for maps;

[0216] According to an embodiment of the present invention, the randomized scenario experiment design specifically includes:

[0217] Select the waiting situation factors, including at least: flight type, travel purpose, waiting time, number of checked baggage, accompanying persons and flight time, and calibrate the corresponding levels, among which:

[0218] The designated flight types are domestic flights and international flights;

[0219] The designated travel purpose is business, leisure travel, or visiting relatives;

[0220] The designated waiting times are 90 minutes, 120 minutes, 150 minutes and 180 minutes;

[0221] The nominal number of checked baggage is 0, 1, 2, or 3 or more;

[0222] The marked flight times are 8:00, 12:30, 17:30, and 22:00;

[0223] A virtual scenario experiment was designed to track and monitor passengers' intention choices in a virtual terminal environment under the premise of given random passenger scenarios, and capture them as a complete passenger activity trajectory.

[0224] Passengers enter the experimental interface. The top of the page shows the flight scenario, which specifies the time and type of flight the passenger will take, the purpose of the trip, the entrance to the airport, the time to arrive at the airport in advance to wait for the flight, the number of accompanying persons, and the number of checked luggage carried on the trip.

[0225] In the center of the page is a search box. Passengers can click the search box to view the list of itinerary points, and click the itinerary point name to select the itinerary point. The selected itinerary point will be highlighted. Passengers can also directly enter the itinerary point name to search, and the searched itinerary point will be highlighted.

[0226] The lower part of the page is a map of the terminal's interior environment, and passengers can select a travel point by clicking on the point on the map;

[0227] Design a randomized virtual experiment. Given a scenario and environment map, passengers pass through a virtual terminal (such as Figure 3 As shown in the figure, you can select the activity point you intend to go to in the environment map, and finally form a complete passenger activity chain (as shown in the figure). Figure 2 shown).

[0228] It should be noted that flight types are divided into four categories: domestic direct flights, domestic transfers, international direct flights, and international transfers. Since the check-in, security check, and boarding gates used by domestic flights and international flights are different, passengers need to choose the corresponding check-in facilities, security check facilities, and boarding gates based on the flight type in the experimental background;

[0229] It should be noted that the random virtual scenarios we assign to passengers include the waiting time of the passengers. The total route time formed by the passengers' choice of activity points cannot exceed the waiting time, otherwise a qualified activity chain cannot be formed.

[0230] According to an embodiment of the present invention, the collection of basic attribute information, historical travel information, travel preference information and activity trajectory information in the airport of passengers specifically includes:

[0231] Passengers' basic attribute information collection: A total of 808 civil aviation passengers' basic attribute information was collected, including: gender, age, annual disposable income, unit type, position and education level;

[0232] Passenger historical travel data collection, including: passenger type, number of civil aviation trips, travel cost ratio, ticket class and whether special civil aviation services are required;

[0233] Passengers’ travel preference data collection, including shopping preferences, dining preferences, beverage preferences and office area preferences;

[0234] Among them, the types of stores preferred for shopping include luxury stores, cosmetics stores, clothing stores, various duty-free stores, specialty stores, home furnishing stores, convenience stores, and leisure parks;

[0235] The types of restaurants preferred by consumers include Chinese fast food, Western fast food, Chinese formal meals, and Western formal meals;

[0236] The types of stores preferred for beverages include milk tea shops, coffee shops, and convenience stores;

[0237] The types of shops preferred by office areas include cafes, restaurants, and rest areas;

[0238] Passengers’ activity trajectory data collection in the airport, a total of 1,783 activity trajectories were collected, including: activity points, activity routes and activity duration;

[0239] According to an embodiment of the present invention, the parameter calibration and specific analysis of civil aviation passenger demand based on the discrete choice model and the formation of preliminary selection probability annotation specifically include:

[0240] Assuming that the set of available store types in the airport faced by passenger n is J(i∈J), the utility function of passenger n choosing store type i is expressed as:

[0241]

[0242] Among them, U in represents the utility of passenger n choosing airport store type i; V in represents the observable utility of passenger n choosing store type i; ε in represents the random utility of traveler n choosing store type i;

[0243] The probability P that traveler n chooses store type i in It is expressed as:

[0244] P in =prob(U in >U jn ;i≠j,j∈J)

[0245] Among them, prob represents the probability function of passenger behavior selection, U in represents the utility of passenger n choosing airport store type i, U jn represents the utility of passenger n choosing airport store type j, where j represents the store type and:

[0246] 0≤P in ≤1

[0247] That is, we get:

[0248]

[0249] Among them, Φ() represents the cumulative distribution function of the standard normal distribution, σ represents the variance value, and ε jn represents the random utility of passenger n choosing store type j, V jn represents the observable utility of traveler n choosing store type j.

[0250] According to an embodiment of the present invention, the construction of a visualized civil aviation passenger portrait knowledge graph specifically includes:

[0251] Segmentally label the collected activity trajectories of civil aviation passengers in the airport environment;

[0252] Process the collected basic passenger attribute information, historical travel information, travel preference information and activity trajectory information in the airport into a triple form of <entity, relationship, entity> or <entity, attribute, attribute value>;

[0253] Import the triple data set into Neo4j to achieve large-scale data storage and build a visual knowledge graph of civil aviation passenger portraits;

[0254] According to an embodiment of the present invention, the segmented annotation of the collected passenger's activity trajectory information in the airport specifically includes:

[0255] The passenger activity trajectories were segmented in chronological order, with each segment representing a different activity stage of airport passengers. A total of 2,943 activity trajectories were collected in chronological order.

[0256] The passenger activity trajectories were segmented by activity points, and the activity sequence, activity distance and activity type of each segment were marked. A total of 8,763 activity trajectories were collected according to activity segmentation data;

[0257] The collected basic attribute information, historical travel information, travel preference information and activity trajectory information of passengers in the airport are processed into triples, including:

[0258] The marked selection probability is used as the probability attribute of the passenger's choice and processed into a triple form of <a passenger, selection probability, probability value>, and a total of 15,334 probability attribute triples are obtained;

[0259] The distance, type and item price of the activity points in the airport environment are taken as activity point attributes and processed into a triple form of <a certain activity point, attribute, attribute value>, and a total of 2174 activity point attribute triple data are obtained;

[0260] The relationship between the link activity point and the store type is processed into a triple form of <a certain activity point, belongs to, a certain store type>, and a total of 812 store type triples are obtained;

[0261] The passenger's activity trajectory is segmented into a triple form of <passenger, choice, activity point>, and a total of 8763 activity trajectory segment triple data are obtained;

[0262] The basic attributes and travel preference information of passengers are marked as personal attributes of passengers and processed into the triple form of <passenger, attribute type, attribute level>. A total of 6251 passenger attribute triple data are obtained.

[0263] According to an embodiment of the present invention, the knowledge graph technology is used to describe and analyze individual and group portraits of civil aviation passengers, such as Figure 4 As shown, specifically including:

[0264] Create a node for each passenger, including their basic attribute characteristics, historical travel characteristics, travel preference characteristics, and activity trajectory characteristics within the airport;

[0265] Define the relationship types between entities, including <select, belong to, attribute, attribute type, selection probability>, define and specify attribute values ​​for each relationship;

[0266] Analyze the common characteristics and behavioral trends of passenger groups through data mining technology, and identify key influencing factors among passenger groups, such as travel purpose, preference type, etc.;

[0267] Leverage Neo4J’s advanced query languages ​​(such as Cypher) to perform complex data queries and pattern recognition, and analyze the correlation between traveler behaviors and preferences, and their impact on travel decisions;

[0268] Use graph visualization tools to extract portrait labels of individual and group passengers and extract valuable insights to support airlines' service improvements, marketing strategies, and operational optimization.

[0269] According to an embodiment of the present invention, providing support for airlines' service improvement, marketing strategy and operation optimization specifically includes:

[0270] Through individual portraits, the airport can gain in-depth understanding of each passenger’s needs and preferences. For example, customized catering services can be provided for passengers with special dietary requirements.

[0271] Based on group portraits, airports can identify target customer groups and develop targeted marketing strategies. For example, for high-end passenger groups, they can provide services such as priority boarding.

[0272] By analyzing passengers’ behaviors and needs within the terminal, airports can more rationally allocate store locations, airport layouts, and other resources, thereby improving overall operational efficiency;

[0273] By analyzing passenger behavior and feedback, airports can identify deficiencies in services and make targeted improvements, thereby improving overall passenger satisfaction;

[0274] According to an embodiment of the present invention, the extraction of portrait labels of individual and group passengers using a graphic visualization tool specifically includes:

[0275] Extract passenger individual portrait labels from the same passenger's choice behavior in different airport environments, and perform visual analysis on the passenger individual portrait;

[0276] Extract passenger individual portrait labels from the selection behaviors of passengers with similar attributes in the same airport environment, and perform visual analysis on passenger individual portraits;

[0277] Analyze the distribution of group passengers from the perspective of different types of stores, extract group portrait labels of civil aviation passengers, and visualize and analyze passenger group portraits;

[0278] Through the passenger group profile provided by this method, the airport can make auxiliary decisions based on the behaviors and preferences in the passenger group profile, arrange the locations of airport stores and other resources, and increase the efficiency of passenger passage;

[0279] The order and location of aircraft take-off and landing can also be adjusted according to the passenger characteristic data in the passenger group portrait, thereby improving the utilization rate of the airport.

[0280] An embodiment of the present invention also provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement a method for characterizing civil aviation passenger portraits based on a knowledge graph as described in the above embodiment.

[0281] The present invention provides an electronic device, comprising: one or more processors, a memory, and one or more computer programs stored in the memory and executable on the processor, wherein the processor executes the program to implement a method for characterizing civil aviation passenger portraits based on a knowledge graph as described in the above embodiment.

[0282] In a specific implementation, the present application provides a computer storage medium and a corresponding data processing unit, wherein the computer storage medium can store a computer program, and when the computer program is executed by the data processing unit, the invention content of a civil aviation passenger profiling method integrating a discrete choice model and a knowledge graph provided by the present invention and some or all of the steps in each embodiment can be run. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0283] Those skilled in the art can clearly understand that the technical solutions in the embodiments of the present invention can be implemented by means of computer programs and their corresponding general hardware platforms. Based on such an understanding, the technical solutions in the embodiments of the present invention can be essentially or partly contributed to the prior art in the form of computer programs, i.e., software products, which can be stored in a storage medium and include several instructions for enabling a device including a data processing unit (which can be a personal computer, a server, a single-chip microcomputer, an MCU or a network device, etc.) to execute the methods described in various embodiments of the present invention or certain parts of the embodiments.

[0284] The present invention provides a method and idea for a civil aviation passenger profiling method that integrates a discrete selection model and a knowledge graph. There are many methods and approaches to implement the technical solution. The above is only a preferred implementation of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention. All components not specified in this embodiment can be implemented using existing technologies.

Claims

1. A civil aviation passenger profiling method integrating discrete choice model and knowledge graph, characterized in that: The specific steps include: Step 1: Determine the entities and relationships in the field of civil aviation passenger portraits, and construct the knowledge ontology of civil aviation passenger portraits in a top-down manner; Step 2: Design a virtual terminal waiting experimental system to collect basic passenger attribute information, historical travel information, travel preference information, and activity trajectory information in the airport while simulating the virtual terminal geographical environment and waiting situation; Step 3: calibrate and analyze the parameters of civil aviation passenger demand based on the discrete choice model, and form a preliminary selection probability label; Step 4: Process the basic attribute information, historical travel information, travel preference information and activity trajectory information in the airport collected in step 2, and construct a visual knowledge graph of civil aviation passenger portraits; Step 5: Use the query language to query and analyze the civil aviation passenger portrait knowledge graph constructed in step 4, and characterize and analyze the individual and group portraits of civil aviation passengers based on the knowledge graph technology; Among them, the design of a terminal virtual waiting experimental system described in step 2 specifically includes: Design an intention survey questionnaire to record passengers' basic attribute information, historical travel information, and historical preference information; Design a virtual terminal passenger waiting experiment, including virtual airport environment design and random situation experiment design; virtual waiting experiment based on behavioral research, randomly assign virtual situation information and airport environment to passengers, and collect passenger activity trajectory information under the changing airport environment; The design of the intention survey questionnaire specifically includes: collecting basic attribute information of passengers, collecting historical travel data of passengers, and collecting historical preference data of passengers; The virtual airport environment design specifically includes: Design the check-in counter in the virtual airport environment according to the check-in counter layout rules, including: determining the area occupied by the check-in island, including the size of the check-in island itself and the width of the surrounding queuing area; determining the layout location of the check-in island, that is, determining the width of the circulation area before and after the check-in counter; determining the layout method of the check-in island, including the layout method of the check-in island can be divided into centralized layout and decentralized layout; Designing a self-service check-in in a virtual environment according to the layout rules of the self-service check-in specifically includes: determining the floor space of the self-service check-in, including the size of the self-service check-in itself and the width of the surrounding queue area; determining the layout location of the self-service check-in, that is, determining the width of the circulation area before and after the self-service check-in counter; determining the layout method of the self-service check-in, including the layout methods of the self-service check-in, which can be divided into centralized layout and decentralized layout; designing the location and layout of the security checkpoint according to the centralized layout / decentralized layout according to the layout rules of the security checkpoint; designing the layout of the non-functional area according to the centralized layout / decentralized layout according to the commercial format ratio of large hub airports at home and abroad and the number of existing large airport stores; The base map rules of the virtual terminal map are designed based on the grid labeling method, and all the unit grids of the map are serialized by assigning a sequence value to each unit grid of the map, specifically including: Assume that the horizontal length x of the global two-dimensional environment max , the vertical length is y max , the side length of each unit grid is a. If it is less than a unit grid, it is processed as a unit grid. Thus, the two-dimensional geographical environment is divided into a series of unit grids to obtain a two-dimensional grid environment model. The total number of unit grids can be expressed as: Among them, n x is the number of unit grid columns, n y is the number of unit grid rows, ceil is the rounding operation; Assign a value to each unit grid. If the i-th unit grid is an obstacle grid, assign a value to the obstacle grid n. i Assign a value of 1; if the jth unit grid is a free grid, then for the free grid n j Assign a value of 0; A rectangular coordinate system is created on the two-dimensional grid environment model, with the upper left corner of the two-dimensional grid environment model as the coordinate origin o, the horizontal direction of the environment model as the x-axis, and the vertical direction of the environment model as the y-axis, and the unit lengths of the two coordinate axes are all integer multiples of the side lengths of the unit grid. Assume that the sequence set is: C = {1, 2, 3, ..., u, u∈N}, if m represents the sequence number, m∈C, and then assign the sequence value m to each grid from top to bottom and from left to right; in the rectangular coordinate system, the rectangular coordinate method and the serial number method correspond to each other. The coordinate position (x m ,y m ), are determined by the sequence number m, as follows: x m =(m-1)mode x +1 Among them, mod is the remainder operation, int is the floor operation, and x m The horizontal coordinate when m is y m is the ordinate when the serial number is m; After calibrating the grids on the map, set the occupied grid number of each facility according to the layout rules; edit the grid attributes, including at least the occupied time, color, mutation and whether it is clickable; Use drawing tools to draw road networks for maps; The randomized scenario experimental design specifically includes: Select waiting situation factors, including at least: flight type, travel purpose, waiting time, number of checked baggage, accompanying persons and flight time, and calibrate corresponding levels for them; Step 3 is to perform parameter calibration and specific analysis on civil aviation passenger demand based on the discrete choice model, and form a preliminary selection probability labeling, which specifically includes: Assuming that the set of available store types in the airport faced by passenger n is J(i∈J), the utility function of passenger n choosing store type i is expressed as: Among them, U in represents the utility of passenger n choosing airport store type i; V in represents the observable utility of passenger n choosing store type i; ε in represents the random utility of traveler n choosing store type i; The probability P that traveler n chooses store type i in It is expressed as: P in =prob(U in >U jn ;i≠j,j∈J) Among them, prob represents the probability function of passenger behavior selection, U in represents the utility of passenger n choosing airport store type i, U jn represents the utility of passenger n choosing airport store type j, where j represents the store type and: 0≤P in ≤1 That is, we get: Among them, Φ() represents the cumulative distribution function of the standard normal distribution, σ represents the variance value, and ε jn represents the random utility of passenger n choosing store type j, V jn represents the observable utility of passenger n choosing store type j, and the probability P in Passengers are labeled as selection probabilities.

2. According to claim 1, a civil aviation passenger profiling method integrating discrete choice model and knowledge graph is characterized in that: The knowledge ontology of the civil aviation passenger portrait map is constructed in a top-down manner as described in step 1, including: Design a first-level ontology framework, including basic passenger attributes, context, and airport environment; Design the secondary ontology, including: The secondary ontology of basic passenger attributes includes: socio-demographic characteristics, travel characteristics, travel preference characteristics and activity chain information. The secondary ontology of basic passenger attributes is further subdivided into attribute entities, including at least: age, gender, occupation, income, position, dietary preferences and shopping preferences; The secondary ontology of the airport environment includes: grids and activity points. Based on the secondary ontology of the airport environment, attribute entities are subdivided downwards, including at least: grid coordinates, grid spacing, activity point attributes, store brands, and store types; The secondary ontology of the context includes: attributes of accompanying persons, flight type, flight departure time, travel purpose, waiting time, and number of checked baggage.

3. According to claim 2, a civil aviation passenger profiling method integrating discrete selection model and knowledge graph is characterized in that: The construction of a visual civil aviation passenger profile knowledge graph described in step 4 specifically includes: Step 4-1, segmentally labeling the passenger's activity trajectory information in the airport collected in step 2; Step 4-2, construct the basic attribute information of passengers, historical travel information, travel preference information and activity trajectory information in the airport collected in step 2 into a triple form of <entity, relationship, entity> or <entity, attribute, attribute value>; Step 4-3, import the above triplet data set into storage, and use visualization tools to build a visual civil aviation passenger portrait knowledge graph.

4. According to claim 3, a civil aviation passenger profiling method integrating discrete choice model and knowledge graph is characterized in that: The description of the knowledge graph technology in step 5 to characterize and analyze the individual and group portraits of civil aviation passengers includes: Step 5-1, create a node in the knowledge graph for each passenger, including the passenger's basic attribute characteristics, historical travel characteristics, travel preference characteristics, and activity trajectory characteristics in the airport; Step 5-2, define the relationship types between entities and specify attributes for each relationship; Step 5-3: Analyze the common characteristics and behavior trends of passenger groups through data mining technology, and identify the key influencing factors among passenger groups; Step 5-4: Use advanced query languages ​​to perform data query and pattern recognition, and analyze the correlation between passenger behaviors and preferences, as well as their impact on travel decisions; Step 5-5, use graph visualization tools to extract portrait labels of individual and group passengers.

5. According to claim 4, a civil aviation passenger profiling method integrating discrete choice model and knowledge graph is characterized in that: Step 5-5 uses a graph visualization tool to extract profile labels for individual and group passengers, including: Extract passenger individual portrait labels from the same passenger's choice behavior in different airport environments, and perform visual analysis on the passenger individual portrait; Extract passenger individual portrait labels from the selection behaviors of passengers with similar attributes in the same airport environment, and perform visual analysis on passenger individual portraits; Analyze the distribution of group passengers from the perspective of different types of stores, extract group portrait labels of civil aviation passengers, and visualize and analyze passenger group portraits; Analyze the distribution trend of group passengers from the perspective of passenger preferences, extract group portrait labels of civil aviation passengers, and visualize and analyze passenger group portraits.

6. According to claim 5, a civil aviation passenger profiling method integrating discrete choice model and knowledge graph is characterized in that: The basic attribute information, historical travel information, travel preference information and activity trajectory information of passengers collected in step 2 specifically include: The basic attribute information of passengers shall be collected, including at least: gender, age, annual disposable income, unit type, position and education level; The passenger's historical travel data collection shall include at least: passenger type, number of civil aviation trips, travel cost ratio, ticket class and whether special civil aviation services are required; The collection of passengers' travel preference data includes at least: shopping preferences, dining preferences, beverage preferences and office area preferences; the above preferences are different store types; The passenger's activity trajectory data collection in the airport includes at least: activity points, activity routes and activity duration.

7. The civil aviation passenger profiling method integrating discrete choice model and knowledge graph according to claim 6 is characterized in that: The passenger's activity trajectory information in the airport collected by segmented annotation in step 4-1 specifically includes: Step 4-1-1, divide the passenger activity trajectory into segments in chronological order, each segment represents a different activity stage of the airport passenger; divide the passenger activity trajectory into segments according to the activity points, and mark the activity sequence, activity distance and activity type of each segment; Step 4-1-2, the collected basic attribute information, historical travel information, travel preference information and activity trajectory information in the airport of the passenger are processed into a triple form, including: According to the selection probability marked in step 3, as the probability attribute of the passenger's selection, it is processed into a triple form of <passenger, selection probability, probability value>; The distance, type and item price of the activity point in the airport environment are taken as activity point attributes and processed into a triple form of <activity point, attribute, attribute value>; The relationship between the link activity point and the store type is processed into the triple form of <activity point, belongs to, store type>; the passenger's activity trajectory is segmented and processed into the triple form of <passenger, choice, activity point>; the passenger's basic attributes and travel preference information are marked as the passenger's personal attributes, which are processed into the triple form of <passenger, attribute type, attribute level>.

Images