Air Ticket Recommendation Method, Device and Electronic Device Based on Large Language Model

Abstract

The present application discloses a flight ticket recommendation method, device, and electronic device based on a large language model, which relate to the fields of artificial intelligence and aviation services. Among them, the method includes: in response to a flight ticket query request initiated by a user, obtaining business scenario prompt words, conversation history, and business information through an agent; obtaining the user's preference information and the status information of each flight route; generating a basic template based on the business scenario prompt words, and extracting knowledge information from the conversation history, business information, the user's preference information, and the status information of each flight route, and filling the knowledge information into the basic template to obtain a target knowledge prompt; processing the flight ticket query request by the large language model according to the target knowledge prompt to generate at least one recommended flight ticket information for the user. The present application solves the technical problems of low accuracy of flight ticket recommendations for users and poor user experience in the prior art.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence and the field of aviation services. Specifically, it relates to a ticket recommendation method, device, and electronic device based on a large language model. Background Art

[0002] With the rapid development of artificial intelligence technology, its applications in all walks of life are becoming increasingly widespread, and the civil aviation sales field is no exception. Traditional civil aviation sales methods mainly rely on ticket data for services. However, this method has gradually revealed its limitations and cannot meet the growing diverse needs of modern passengers.

[0003] Currently, the sales models of most airlines mainly focus on providing consumers with basic ticket information, such as the specific departure and arrival times of flights, ticket prices, and other basic elements, in order to provide purchase options for users. However, this traditional service method has an obvious defect, that is, it does not comprehensively consider the diverse needs of users during the ticket purchase process. In fact, the user's ticket purchase decision is a complex process influenced by many factors, including the user's personal business behavior patterns (such as business trips or private travel), personal preference settings (such as preferring window seats or aisle seats, having special meal requirements, etc.), and real-time changing environmental factors (such as weather conditions, traffic congestion, emergencies at the destination, etc.). These factors act together on the user's decision-making process and play a crucial role. Unfortunately, the existing sales models have not effectively integrated the above key information into the service, resulting in users lacking sufficient information support when purchasing tickets and being difficult to obtain the choice that best suits their actual needs. This situation not only seriously affects the user's ticket purchase experience, making them feel inconvenient and dissatisfied, but also weakens the airline's ability to provide personalized and high-quality services, restricting the further improvement of its market competitiveness and service level.

[0004] In summary, in the face of the rapid development of artificial intelligence technology and the growing diverse needs in the civil aviation sales field, the existing traditional sales models have become inadequate. This model is limited to providing basic ticket information and fails to fully consider the diverse factors of users, resulting in users lacking comprehensive information support when making ticket purchase decisions. This not only affects the user's ticket purchase experience but also leads to a low accuracy of ticket recommendations for airlines to users.

[0005] In response to the above problems, no effective solutions have been proposed yet. Summary of the Invention

[0006] The present application provides a flight ticket recommendation method, apparatus, and electronic device based on a large language model, so as to at least solve the technical problems of low accuracy of flight ticket recommendations for users and poor user experience caused by only referring to the departure point, destination, and price of the current flight route in the existing flight ticket recommendation process.

[0007] According to one aspect of the present application, there is provided a flight ticket recommendation method based on a large language model, including: in response to a flight ticket query request initiated by a user, obtaining, by an agent, business scenario prompt words, conversation history, and business information, where the business scenario prompt words are used to represent the business scenario information corresponding to the agent, the conversation history is the historical interaction information between the user and the software system, and the business information is information related to the aviation business obtained from application services other than the agent according to the flight ticket query request and the business scenario information corresponding to the agent; obtaining the user's preference information and the status information of each flight route, where the preference information at least includes the user's habitual information when handling aviation business and the user's behavioral preference information when selecting air travel; generating a basic template based on the business scenario prompt words, extracting knowledge information from the conversation history, business information, the user's preference information, and the status information of each flight route, and filling the knowledge information into the basic template to obtain a target knowledge prompt; processing the flight ticket query request by the large language model based on the target knowledge prompt to generate at least one recommended flight ticket information for the user.

[0008] Optionally, obtaining the user's preference information and the status information of each flight route includes: determining, by means of keyword recognition, the user's habitual information when handling aviation business and the user's behavioral preference information when selecting air travel from the conversation history and the user's historical itinerary information to obtain the user's preference information; determining the ticket price information, flight punctuality rate, flight delay risk information, and weather information along the route of each flight route from the business information; and using the ticket price information, flight punctuality rate, flight delay risk information, and weather information of each flight route as the status information of each flight route.

[0009] Optionally, generating a basic template based on the business scenario prompt words includes: obtaining a first type of prompt words, a second type of prompt words, and a third type of prompt words, where the first type of prompt words is used to prompt the parsing method adopted by the large language model when parsing the user's intention, the second type of prompt words is used to prompt the reasoning angle selected by the large language model when processing the target knowledge prompt, and the third type of prompt words is used to constrain the data format filled into the basic template; and generating a basic template according to the first type of prompt words, the second type of prompt words, the third type of prompt words, and the business scenario prompt words.

[0010] Optionally, knowledge information is extracted from the conversation history, business information, user preference information, and status information of each flight route, including: extracting corresponding knowledge information from each type of information in the conversation history, business information, user preference information, and status information of each flight route using the same aviation business perspective; and / or extracting corresponding knowledge information from each type of information in the conversation history, business information, user preference information, and status information of each flight route using different aviation business perspectives.

[0011] Optionally, after extracting knowledge information from the conversation history, business information, user preference information, and status information of each flight route, the air ticket recommendation method based on the large language model further includes: performing semantic splicing processing on at least two pieces of extracted knowledge information according to the semantic content corresponding to the knowledge information to obtain new knowledge information.

[0012] Optionally, filling the knowledge information into a basic template to obtain a target knowledge prompt, including: screening the knowledge information extracted from the conversation history, business information, user preference information, and status information of each flight route according to the business scenario prompt words in the basic template to obtain target content with a correlation coefficient greater than a preset threshold with the business scenario prompt words; filling the target content into the basic template to obtain the target knowledge prompt.

[0013] Optionally, filling the knowledge information into a basic template to obtain a target knowledge prompt, including: screening the knowledge information extracted from the business information according to the business scenario prompt words in the basic template to obtain first content with a correlation coefficient greater than a preset threshold with the business scenario prompt words in the knowledge information; integrating the business scenario prompt words and the first content into second content according to the basic template, and screening the knowledge information extracted from the conversation history according to the second content to obtain third content with a correlation coefficient greater than a preset threshold with the second content in the knowledge information; integrating the second content and the third content into fourth content according to the basic template, and screening the knowledge information extracted from the user preference information and the status information of each flight route according to the fourth content to obtain fifth content with a correlation coefficient greater than a preset threshold with the fourth content in the knowledge information; integrating the fourth content and the fifth content and filling them into the basic template to obtain the target knowledge prompt.

[0014] Optionally, the large language model processes the flight ticket query request according to the target knowledge prompt to generate at least one recommended flight ticket information for the user, including: the large language model performs intent parsing and flight ticket information reasoning on the target knowledge prompt and the flight ticket query request according to the first type of prompt words and the second type of prompt words, and determines at least one recommended flight ticket information that meets the user's intent according to the predicted user intent and flight ticket information reasoning result, where the presentation form of the recommended flight ticket information includes at least one of text form, voice form, image form, and video form.

[0015] Optionally, the flight ticket recommendation method based on the large language model further includes: when the presentation form of the recommended flight ticket information is in voice form, the large language model determines the language, tone, and voice style to be used when communicating the recommended flight ticket information to the user through voice according to the target knowledge prompt; when the presentation form of the recommended flight ticket information is in text form, the large language model determines the character form and text expression style to be used when communicating the recommended flight ticket information to the user through text according to the target knowledge prompt; when the presentation form of the recommended flight ticket information is in image form, the large language model determines the image elements and the combination method of the image elements to be used when communicating the recommended flight ticket information to the user through image according to the target knowledge prompt; when the presentation form of the recommended flight ticket information is in video form, the large language model determines the audio elements and video elements to be used when communicating the recommended flight ticket information to the user through video according to the target knowledge prompt.

[0016] According to another aspect of the present application, there is also provided a flight ticket recommendation device based on a large language model, including: a first acquisition unit, configured to, in response to a flight ticket query request initiated by a user, acquire a business scenario prompt word, a conversation history, and business information through an intelligent agent, where the business scenario prompt word is used to represent the business scenario information corresponding to the intelligent agent, the conversation history is the historical interaction information between the user and the software system, and the business information is aviation-related information obtained from application services other than the intelligent agent according to the flight ticket query request and the business scenario information corresponding to the intelligent agent; a second acquisition unit, configured to acquire the user's preference information and the status information of each flight route, where the preference information at least includes the user's habitual information when handling aviation business and the user's behavioral preference information when selecting air travel; a first processing unit, configured to generate a basic template according to the business scenario prompt word, extract knowledge information from the conversation history, business information, the user's preference information, and the status information of each flight route, and fill the knowledge information into the basic template to obtain a target knowledge prompt; a second processing unit, configured to process the flight ticket query request through the large language model according to the target knowledge prompt to generate at least one recommended flight ticket information for the user.

[0017] According to another aspect of the present application, there is also provided a computer-readable storage medium, wherein the computer-readable storage medium includes a stored executable program, and when the executable program runs, it controls the device where the computer-readable storage medium is located to execute the above-mentioned air ticket recommendation method based on a large language model.

[0018] According to another aspect of the present application, there is also provided an electronic device, including: a memory storing an executable program; a processor for running the program, and when the program runs, it executes the above-mentioned air ticket recommendation method based on a large language model.

[0019] According to another aspect of the present application, there is also provided a computer program product, including computer instructions, characterized in that when the computer instructions are executed by a processor, the steps of the above-mentioned air ticket recommendation method based on a large language model are implemented.

[0020] In the present application, first, in response to an air ticket query request initiated by a user, a business scenario prompt word, a conversation history, and business information are obtained through an agent, where the business scenario prompt word is used to represent the business scenario information corresponding to the agent, the conversation history is the historical interaction information between the user and the software system, and the business information is information related to the aviation business obtained from application services other than the agent according to the air ticket query request and the business scenario information corresponding to the agent. Then, the preference information of the user and the status information of each flight route are obtained, where the preference information at least includes the habit information of the user when handling aviation business and the behavior preference information of the user when choosing air travel. Subsequently, a basic template is generated based on the business scenario prompt word, and knowledge information is extracted from the conversation history, business information, the preference information of the user, and the status information of each flight route, and the knowledge information is filled into the basic template to obtain a target knowledge prompt word. Finally, the air ticket query request is processed by a large language model according to the target knowledge prompt word to generate at least one recommended air ticket information for the user.

[0021] As can be seen from the above, first of all, through the comprehensive application of business scenario prompt words, conversation history, and business information, the intelligent agent can deeply understand the specific needs and preferences of users, rather than just making recommendations based on the origin, destination, and price. For example, users may have habits such as preferring flights of a specific airline, preferring specific departure or arrival times, having special meal requirements, etc., as well as behavioral preference information such as weather preferences and emphasis on flight punctuality. Through the processing of the large model, the intelligent agent can accurately capture and apply these preferences, significantly improving the personalization level of flight ticket recommendations. Secondly, in this application, the intelligent agent not only pays attention to user preferences but also obtains the status information of each route in real time, including but not limited to flight punctuality, aircraft type information, cabin status, fare fluctuations, and early warnings of emergencies. This means that when recommending flight tickets, the system can comprehensively consider various factors, provide more comprehensive and accurate flight information, and help users make more reasonable itinerary choices.

[0022] Furthermore, in this application, the business scenario prompt words are used to generate a basic template, and then the intelligent agent combines the conversation history, business information, user preferences, and route status to extract key knowledge information and fill it into the template to form a target knowledge prompt. This design ensures that when the large model processes flight ticket query requests, it can focus on using this knowledge information to optimize the reasoning process and provide recommendation results that are closer to user needs. Finally, through the in-depth understanding and reasoning of the large model, the intelligent agent can optimize the processing of flight ticket query requests. The generated recommendation information is not only based on factual data but also integrates intelligent analysis, and can provide the best flight ticket recommendation solutions from multiple dimensions such as the user's historical behavior, current environmental conditions, and business requirements.

[0023] Thus, through the application of the above technical features, this application significantly improves the accuracy and personalization level of flight ticket recommendations, enabling users to obtain recommendation results that not only meet their actual needs but also have a high degree of relevance when querying flight information. This not only simplifies the user's decision-making process but also significantly improves user satisfaction and experience, providing users with a one-stop intelligent civil aviation sales service and making the user's travel planning more convenient and comfortable. The application of the technical features of this application solves the problems of low accuracy and poor user experience caused by only considering basic elements in flight ticket recommendations in the prior art. By deeply mining user preferences, integrating route status information, and applying the wisdom of the large model, it realizes the personalized and intelligent upgrade of flight ticket recommendation services, greatly improving the ticket purchase experience and satisfaction of users in the civil aviation sales field. Brief Description of the Drawings

[0024] The drawings described herein are used to provide a further understanding of this application and form a part of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation of this application. In the drawings:

[0025] Figure 1 The hardware structure block diagram of a computer terminal for implementing an air ticket recommendation method based on a large language model is shown;

[0026] Figure 2 It is a flowchart of an optional air ticket recommendation method based on a large language model according to an embodiment of the present application;

[0027] Figure 3 It is a schematic diagram of the working process of an optional agent according to an embodiment of the present application;

[0028] Figure 4 It is a schematic diagram of an optional air ticket recommendation device based on a large language model according to an embodiment of the present application. Detailed implementation manners

[0029] In order to enable those skilled in the art of the present technology to better understand the solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present application described here can be implemented in an order different from those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0031] It should also be noted that the information collected in this application (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) are information and data that have been authorized by the user or fully authorized by all parties. Moreover, for the processing of relevant data such as collection, storage, use, processing, transmission, provision, disclosure, and application, all comply with relevant laws, regulations, and standards, necessary confidentiality measures are taken, it does not violate public order and good customs, and corresponding operation entrances are provided for users to choose to authorize or refuse. For example, interfaces are set between this system and relevant users or institutions to provide corresponding operation entrances for users to choose to agree or refuse the results of automated decision-making; if the user chooses to refuse, the expert decision-making process will be entered.

[0032] According to an embodiment of the present application, an embodiment of a method for recommending air tickets based on a large language model is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0033] The method embodiment provided by the embodiment of the present application can be executed on a mobile terminal, a computer terminal, or a similar computing device. Figure 1 A hardware structure block diagram of a computer terminal (or mobile device) for implementing the method for recommending air tickets based on a large language model according to the embodiment of the present application is shown. As Figure 1 shown, the computer terminal 10 (or mobile device) may include one or more processors 102 (the processors 102 may include, but are not limited to, processing devices such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may further include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply, and / or a camera. Those of ordinary skill in the art can understand that Figure 1 the structure shown is only schematic and does not limit the structure of the above-mentioned electronic device. For example, the computer terminal 10 may further include more or fewer components than Figure 1 shown, or have a different configuration from Figure 1 shown.

[0034] It should be noted that one or more of the above-mentioned processors 102 and / or other data processing circuits can generally be referred to as "data processing circuits" herein. The data processing circuit can be embodied in software, hardware, firmware, or any combination thereof, in whole or in part. In addition, the data processing circuit can be a single independent processing module, or integrated in whole or in part into any one of other components in the computer terminal 10 (or mobile device). As the large language model-based air ticket recommendation method involved in the embodiments of the present application, the data processing circuit is a kind of processor control (such as the selection of a variable resistance terminal path connected to an interface).

[0035] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the large language model-based air ticket recommendation method in the embodiments of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, implements the above-mentioned large language model-based air ticket recommendation method. The memory 104 can include high-speed random access memory, and can also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memories. In some instances, the memory 104 can further include a memory remotely set relative to the processor 102, and these remote memories can be connected to the computer terminal 10 through a network. Examples of the above-mentioned network include but are not limited to the Internet, enterprise intranet, local area network, mobile communication network, and their combinations.

[0036] The transmission device 106 is used to receive or send data via a network. Specific examples of the above-mentioned network can include the wireless network provided by the communication provider of the computer terminal 10. In one instance, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station and thus communicate with the Internet. In one instance, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0037] The display can be, for example, a touch-screen liquid crystal display (LCD), which enables the user to interact with the user interface of the computer terminal 10 (or mobile device).

[0038] Under the above operating environment, the present application provides a large language model-based air ticket recommendation method as Figure 2 shown. Figure 2 It is a flowchart of an optional large language model-based air ticket recommendation method according to an embodiment of the present application. As Figure 2 shown, the method includes the following steps:

[0039] Step S201, in response to a ticket query request initiated by a user, obtain a business scenario prompt word, a conversation history, and business information through an agent.

[0040] The business scenario prompt word is used to represent the business scenario information corresponding to the agent. The conversation history is the historical interaction information between the user and the software system. The business information is information related to the aviation business obtained from application services other than the agent according to the ticket query request and the business scenario information corresponding to the agent.

[0041] Optionally, the business scenario prompt word is a key information carrier when the agent interacts with the large language model, and is used to clarify the current business environment and task requirements of the agent. In this application, for a ticket query request, the business scenario prompt word specifically represents the business scenario information of air ticket sales, such as key elements like the destination of the query, departure time, cabin preference, airline preference, etc. Through these prompt words, the large language model can quickly understand the background of the request and the business goals of the agent, so that when processing the request, it can more specifically call and integrate relevant knowledge information to provide more personalized and accurate flight recommendations for users.

[0042] Secondly, the conversation history refers to all interaction records between the user and the software system before the current conversation. In the context of this application, the conversation history includes but is not limited to historical records of operations such as ticket queries, bookings, and itinerary changes made by the user in the civil aviation sales software, as well as specific requirements and preferences expressed by the user in the conversation. By analyzing the conversation history, the agent can identify the user's past behavior patterns, such as preferred airlines, frequently flown routes, habitual fare ranges, and whether the user often travels with pets. This information is crucial for building a user profile, predicting user preferences, and providing personalized services.

[0043] In addition, the business information refers to information related to the aviation business provided by other application services other than the agent itself. The business information covers various external data required in air ticket sales, such as: real-time flight schedules, fare information, airline policies, weather forecasts, traffic conditions, flight punctuality rate statistics, seasonal fare change trends, preferential information, and cabin information for specific routes, etc. The acquisition and integration of business information are the basis for the agent to provide accurate and comprehensive services, ensuring that when processing a ticket query request, the agent can obtain the latest and most complete aviation business data, so as to generate the best flight recommendations for users.

[0044] It should be noted that an agent refers to an automated component that performs specific tasks in a system. It can autonomously interact with users and process business requests according to preset rules and logics. In the civil aviation sales scenario of this application, the agent is designed as a business expert agent, specifically responsible for ticket query and recommendation services. It conducts in-depth reasoning through a large language model based on business scenario prompt words, conversation history, business information, as well as user preferences and route status, and generates personalized and intelligent ticket recommendation suggestions.

[0045] Step S202: Obtain the user's preference information and the status information of each route.

[0046] In step S202, the preference information at least includes the user's habitual information when handling aviation business and the user's behavioral preference information when choosing air travel.

[0047] Optionally, the preference information refers to the personal tendencies or choice preferences expressed or implied by the user when conducting aviation business, including the user's habitual information when handling aviation business and the user's behavioral preference information when choosing air travel. Among them, the acquisition and application of preference information are crucial for enhancing the personalization and intelligence of civil aviation sales services.

[0048] Optionally, the user's habitual information when handling aviation business refers to the fixed patterns or preferences formed by the user when using civil aviation sales software, such as the user's tendency to choose a specific airline, preference for morning or evening flights, frequent choice of business class or economy class, whether there are special meal requirements (vegetarian, gluten-free, etc.), whether wheelchair service is needed, and the user's trade-off preference between ticket price and service. These habitual information are recognized and integrated by the agent through historical conversation records and business operation historical data, and are used to construct a more accurate user profile.

[0049] Optionally, the user's behavioral preference information when choosing air travel involves the user's choice tendency during the flight decision-making process. For example, the user may be more inclined to choose a route with a shorter transfer time, prefer direct flights rather than transfers, have a higher requirement for flight punctuality, or have a special preference for in-flight entertainment facilities. By analyzing the user's query patterns and selection behaviors, the agent can understand the user's preferences when choosing air travel, and thus make more accurate judgments in ticket recommendations.

[0050] In addition, the route status information refers to all dynamic information related to a specific route. The route status information reflects the latest status of the route at the query moment, and is crucial for the user to make a ticket purchase decision. In this application, the route status information obtained by the agent should at least include at least one of the following information:

[0051] Real-time flight schedule: This includes the estimated departure and arrival times of flights, which is very important for trip planning.

[0052] Fare information: Details the fares for different cabins and whether there are any special promotions or discounts.

[0053] Flight punctuality rate: Through statistical analysis, provides the average delay time of flights or the probability of on-time departure, helping users evaluate the reliability of flights.

[0054] Weather forecast: Checks the weather conditions at the departure and destination airports of the flight, which is crucial for itinerary arrangements sensitive to weather.

[0055] Traffic conditions: Understands the traffic congestion near the airport to facilitate users in planning the time to arrive at the airport.

[0056] Cabin information: Finds out the remaining seats in different cabins on the flight to provide seat selections that better meet users' needs.

[0057] Service policies: Include baggage allowance regulations, pet policies, availability of special meals, etc., to meet users' specific needs.

[0058] Disaster warnings: In cases where specific flight routes may be affected by natural disasters (such as typhoons, volcanic eruptions, etc.), the system should provide warning information to help users avoid risks.

[0059] By obtaining preference information and route status information, the intelligent agent can provide users with more personalized and comprehensive flight ticket query and recommendation services. Preference information enables the system to understand users' specific needs and inclinations, while route status information ensures that the recommended flight information is up-to-date and accurate, thus improving service satisfaction and user experience. Through the intelligent reasoning of the large model, the intelligent agent can comprehensively analyze this information to generate the optimal flight ticket recommendations for individual users, effectively solving the problem of over-generalized flight ticket recommendations and poor user experience in the existing technology.

[0060] Step S203: Generate a basic template according to the business scenario prompt words, extract knowledge information from the conversation history, business information, users' preference information, and the status information of each route, and fill the knowledge information into the basic template to obtain the target knowledge prompt words.

[0061] Optionally, the business scenario prompt words are the starting point for the interaction between the intelligent agent and the large model, and the business scenario prompt words precisely describe the current business scenario and requirements. In the scenario of flight ticket query, the business scenario prompt words may include specific prompts such as "query international flights", "family travel", "carry pets", "business travelers", etc. These prompt words guide the large model to understand the specific business environment and goals of the intelligent agent, so as to generate relevant knowledge information more accurately.

[0062] The basic template is a structured framework generated under the guidance of business scenario prompts and is used for subsequent filling and integration of knowledge information. The design of the basic template should be closely related to the business scenario prompts. For example, in the flight ticket query scenario, the template may include fields such as flight information, price information, route preferences, and special service requirements, which will be filled in subsequent steps to meet the specific needs of users.

[0063] Optionally, the knowledge information is useful information comprehensively analyzed by the agent in this application based on business scenario prompts, conversation history, business information, user preferences, and route status information. The extraction and integration of knowledge information is an important part of the agent's work. It includes the user's personalized needs (such as preferring direct flights rather than transfers, sensitivity to ticket prices, and special service requirements), real-time route status (such as flight on-time rate, weather conditions, and seat availability), and historical behavior patterns (such as frequent flight routes and price range preferences). The accuracy and comprehensiveness of the knowledge information directly determine the quality of flight ticket recommendations.

[0064] Optionally, the target knowledge prompt is an instruction statement generated after filling the extracted knowledge information into the basic template and is used to guide the intelligent reasoning of the large model. This prompt integrates information from multiple aspects such as business scenarios, user preferences, and route status, enabling the large model to fully consider these factors when processing flight ticket query requests and generate more personalized and user-demand-compliant flight ticket recommendation information. The generation of the target knowledge prompt is the core of the interaction between the agent and the large model. The target knowledge prompt not only contains specific business data but also implies the agent's understanding of user needs and judgment of business scenarios, which is the key to intelligent services.

[0065] For example, when the user is a frequent business traveler who prefers direct flights and has high requirements for flight on-time rates, the agent will identify the user's business traveler identity from the conversation history, discover the user's preference for direct flights from the user's query records, and obtain the on-time rate data of the current route through business information. At the same time, the agent will extract weather forecasts and traffic conditions from the route status information to ensure that the recommended flights are not affected by bad weather or traffic. Based on this information, the basic template generated by the agent may include fields such as "business traveler preferences", "direct flights", "high on-time rate requirements", and "no interference from bad weather". Subsequently, the agent fills this knowledge information into the template to generate a target knowledge prompt, such as: "Please recommend direct flights for business travelers, giving priority to flights with a high on-time rate, ensuring that the flights are not affected by bad weather, and at the same time providing the flight ticket price information and available cabin classes."

[0066] Through this process, the target knowledge prompt received by the large model contains a comprehensive description of the user's needs. It can perform intelligent reasoning based on the user's specific preferences and the status of the current flight route, and finally generate personalized flight ticket recommendations for the business traveler. This process fully demonstrates the important value of this application in enhancing service personalization and intelligence.

[0067] Step S204: Use the large language model to process the flight ticket query request based on the target knowledge prompt, and generate at least one recommended flight ticket information for the user.

[0068] Optionally, the large language model is a complex, deep learning-based artificial intelligence model that has been trained on a large amount of text data and can understand and generate high-quality natural language text. In this application, the large language model is used as the core reasoning component, responsible for parsing and understanding the context information in the target knowledge prompt, including user preferences, flight route status, business scenario descriptions, etc., and then performing intelligent reasoning based on this information to generate flight ticket recommendation information that better meets the user's needs. The strength of the large language model lies in its ability to capture the nuances of language and perform complex logical reasoning based on this information, such as understanding the user's preference for direct flights and evaluating the impact of weather conditions on flight selection, so as to provide more accurate and user-friendly services.

[0069] The recommended flight ticket information is the flight ticket options generated by the agent through the large language model processing the target knowledge prompt, which is tailored to the specific needs of the user. This is not just a simple list of prices and times, but contains comprehensive information on the user's personalized needs, such as the on-time rate of the flight, cabin service, price, airline evaluation, and descriptions of the impact of weather and traffic conditions on the flight. The generation of the recommended flight ticket information takes into account the specific preferences expressed by the user in the conversation history, the real-time requirements for flight status, and the knowledge information refined from the business scenario prompt words and flight route status information. The agent ensures that each piece of recommended flight ticket information can meet the specific needs of the user in the current scenario and provide the best travel option.

[0070] For example, when processing a user's flight ticket query request, the target knowledge prompt might describe: "The user prefers morning flights, has a high requirement for the on-time rate of the flight, and hopes to have extra legroom." After receiving this prompt, the large language model will search the database for flights that meet the conditions based on the user's preferences and flight route status information. It will give priority to flights departing in the morning with a high on-time rate and filter out the cabins that provide extra legroom service. Finally, the large language model will generate one or more recommended flight ticket information, each of which will detail the flight time, on-time rate, cabin details (including options for extra legroom), price, and airline policies, ensuring that the user can make the most suitable decision based on this information.

[0071] With this technical feature, the agent of the present application can generate highly personalized flight ticket recommendation information by virtue of the intelligent reasoning ability of the large language model, thus significantly improving the user experience and satisfaction of civil aviation sales services. This reasoning mechanism based on knowledge prompts can not only accurately capture the personalized needs of users, but also flexibly respond to the changing route status, making the flight ticket recommendation service more intelligent, efficient and user-friendly.

[0072] Based on the content of the above steps S201 to S204, first of all, through the comprehensive application of business scenario prompts, conversation history and business information, the agent of the present application can deeply understand the specific needs and preferences of users, rather than just making recommendations based on the departure point, destination and price. For example, users may have habits such as preferring flights of a specific airline, preferring specific departure or arrival times, having special meal requirements, etc., as well as behavioral preference information such as weather preferences and emphasis on flight punctuality. Through the processing of the large model, the agent can accurately capture and apply these preferences, significantly improving the personalization level of flight ticket recommendations. Secondly, in the present application, the agent not only pays attention to user preferences, but also obtains the status information of each route in real time, including but not limited to flight punctuality, aircraft type information, cabin status, fare fluctuations and emergency warnings. This means that when recommending flight tickets, the system can comprehensively consider various factors, provide more comprehensive and accurate flight information, and help users make more reasonable itinerary choices.

[0073] Furthermore, in the present application, the business scenario prompt is used to generate a basic template, and the agent then combines the conversation history, business information, user preferences and route status to extract key knowledge information and fill it into the template to form a target knowledge prompt. This design ensures that when the large model processes flight ticket query requests, it can focus on using this knowledge information, optimize the reasoning process, and provide recommendation results closer to user needs. Finally, through the in-depth understanding and reasoning of the large model, the agent can optimize the processing of flight ticket query requests, and the generated recommendation information is not only based on factual data, but also integrates intelligent analysis, and can provide the best flight ticket recommendation plan from multiple dimensions such as the user's historical behavior, current environmental conditions, and business needs.

[0074] It can be seen that through the application of the above technical features, this application significantly improves the accuracy and personalization level of flight ticket recommendations. When users query flight information, they can obtain recommendation results that not only meet their actual needs but also have a high degree of relevance. This not only simplifies the user's decision-making process but also significantly improves user satisfaction and experience, providing users with a one-stop intelligent civil aviation sales service and making the travel planning more convenient and comfortable for users. The application of the technical features of this application solves the problems of low accuracy and poor user experience in flight ticket recommendations in the prior art that only consider basic elements. By deeply mining user preferences, integrating route status information, and applying the wisdom of large models, it realizes the personalized and intelligent upgrade of flight ticket recommendation services, greatly improving the ticket purchase experience and satisfaction of users in the field of civil aviation sales.

[0075] In an optional embodiment, in order to obtain the user's preference information and the status information of each route, the intelligent agent can, through keyword recognition, determine the user's habitual information when handling air business and the user's behavioral preference information when choosing air travel from the conversation history and the user's historical itinerary information, so as to obtain the user's preference information; determine the ticket price information, flight punctuality rate, flight delay risk information, and weather information along the route of each route from the business information. Finally, the intelligent agent takes the ticket price information, flight punctuality rate, flight delay risk information, and weather information of each route as the status information of each route.

[0076] Optionally, keyword recognition is the basic technology for the intelligent agent to analyze the user's conversation history and historical itinerary information to determine the user's habitual and behavioral preference information. The intelligent agent identifies keywords related to flight queries, reservations, and preferences by analyzing the text or voice conversations input by the user. For example, if the user mentions information such as "I usually choose business class", "I like direct flights and avoid transfers", "I need wheelchair service" in past conversations, the intelligent agent can, through keyword recognition, extract the user's preference for cabin class, preference for flight type, and special needs in air services. These information together constitute a part of the user's preference information. The accuracy and efficiency of keyword recognition are the key to the intelligent recommendation ability of the entire system, which can ensure that the intelligent agent extracts the most relevant information points from historical data.

[0077] By obtaining user preferences through keyword recognition and obtaining the route status in real time from external business information, the intelligent agent can generate highly personalized and intelligent flight ticket recommendation information in the process of processing flight ticket query requests. This recommendation mechanism not only considers the user's personal preferences but also fully considers the actual status of flights, thus greatly improving the user experience and the intelligent level of civil aviation sales services.

[0078] For example, the agent discovers through keyword recognition that the user prefers business class and values flight punctuality. Then, by combining the route status information obtained from business information (such as there are remaining business class seats on a direct flight, with a punctuality rate as high as 98% and clear weather along the way), these information are integrated into route status information, and then a recommendation message is generated: "Based on your preferences, we recommend the business class seats on a certain airline at 8 pm next Tuesday. This flight has a high punctuality rate and good weather conditions along the way, which is very suitable for your travel needs." This way of recommendation not only reflects the personalization of the service but also ensures the timeliness and accuracy of the information, which is a specific manifestation of the technical features of the agent in this application in practical applications.

[0079] Through the application of the above technical features, the agent can provide users with more considerate, accurate, and timely ticket recommendations than traditional civil aviation sales services, thereby enhancing the user's ticket purchasing experience.

[0080] In an alternative embodiment, the agent can obtain the first type of prompt words, the second type of prompt words, and the third type of prompt words. Among them, the first type of prompt words is used to prompt the parsing method adopted by the large language model when parsing the user's intention, the second type of prompt words is used to prompt the inference angle selected by the large language model when processing the target knowledge prompt, and the third type of prompt words is used to constrain the data format filled into the basic template. Then, the agent generates a basic template according to the first type of prompt words, the second type of prompt words, the third type of prompt words, and the business scenario prompt words.

[0081] Optionally, the first type of prompt words are directive words or phrases used to guide the large language model on how to understand the user's intention when the agent interacts with the large language model. These prompt words can include, but are not limited to, parsing the user's emotion, identifying the urgency of the user's needs, understanding the business type of the user's query, etc. For example, when processing the user's flight query request, the first type of prompt words may be: "Please identify whether the user prefers morning flights", "Determine whether the user is asking about the flight delay compensation policy", "Analyze whether the user is sensitive to ticket prices", etc. Through these prompts, the large model can more accurately understand the user's intention and needs in a specific conversation, so as to generate response content that better meets the user's needs.

[0082] The second type of prompting words is to ensure that when the large language model processes the target knowledge prompting words, it can reason and integrate information from the correct perspective or dimension. These prompting words are usually related to specific business scenarios, helping the large model focus on the most relevant information points and reasoning directions. For example, in the flight ticket recommendation scenario, the second type of prompting words may include: "Optimize flight recommendations from the price perspective", "Consider flight punctuality rates for recommendation ranking", "Provide personalized recommendation reasons based on user behavior preferences", etc. These prompting words ensure that when the large model processes flight query requests, it can comprehensively consider multiple dimensions such as ticket price, punctuality rate, and user preferences, and provide the most reasonable and personalized recommendation plan for users.

[0083] The third type of prompting words aims to standardize the data format when the large language model generates response information, ensuring the structuring and readability of the information. Such prompting words can be to require the large model to present information in a specific structure (such as a table, list, etc.), use specific terms or abbreviations, or follow a specific logical order, etc. For example, if the target knowledge prompting words generated by the agent require displaying the detailed information of each flight, the third type of prompting words may be: "Please list the departure and arrival times, ticket prices, and cabin class information of all flights in a table form", "Use airline codes instead of full names", etc. These format constraints make the response information generated by the large model clearer and more organized, facilitating subsequent processing by the agent and presentation to users.

[0084] Then, the agent generates a basic template based on the first type of prompting words, the second type of prompting words, the third type of prompting words, and the business scenario prompting words.

[0085] By comprehensively applying these three types of prompting words, the agent can ensure that when the large language model processes requests, it can not only accurately understand the user's needs, but also start from the correct reasoning perspective, while ensuring the structuring and standardization of information presentation, so as to provide users with efficient, accurate, and easy-to-understand civil aviation sales services. Through this mechanism, the interaction between the agent and the large language model becomes more intelligent and customized, significantly improving the overall quality of the service and the user experience.

[0086] In an optional embodiment, in order to extract knowledge information from the conversation history, business information, user preference information, and status information of each flight route, the agent can use the same aviation business perspective to extract corresponding knowledge information from each type of information in the conversation history, business information, user preference information, and status information of each flight route; and / or, use different aviation business perspectives to extract corresponding knowledge information from each type of information in the conversation history, business information, user preference information, and status information of each flight route.

[0087] Optionally, the agent can extract knowledge information using the same aviation business perspective. When the agent extracts knowledge from information from different sources using the same aviation business perspective, it means that the agent will uniformly process and integrate data from the conversation history, business information, user preferences, and route status around a specific business requirement or standard. For example, in a flight query scenario, the agent may select "on-time rate" as the unified business perspective, identify from the conversation history whether the user has ever expressed dissatisfaction with flight delays, obtain the on-time rate data of the current route from the business information, analyze from the user preference information whether the user tends to choose flights with a high on-time rate, and monitor the actual on-time rate from the route status information. From this perspective, the agent can synthesize all relevant information to generate knowledge information about the flight on-time rate, such as "The user prefers flights with a high on-time rate. The average on-time rate of the current route is 95%. The user mentioned dissatisfaction with delayed flights in the historical conversation. It is recommended to give priority to recommending flights with a high on-time rate."

[0088] On the other hand, the agent can also extract knowledge from various information sources according to different aviation business perspectives to meet the user's needs in multiple dimensions. For example, in a flight query scenario, the agent may use "price", "service", and "convenience" as different business perspectives, identify from the conversation history whether the user is price-sensitive, analyze from the user preference information whether the user prefers business class service, and obtain from the route status information whether the departure and arrival times of the flight meet the user's time window requirements. In this way, the agent can generate multi-dimensional knowledge information, such as "The user prefers economy class fares, business class service, and hopes that the flight departure and arrival times are in the afternoon on weekdays", so that when recommending flights, the agent can comprehensively consider the user's multiple needs and provide more personalized and accurate services.

[0089] Whether using the same aviation business perspective or different business perspectives to extract knowledge information, the purpose of the agent is to better understand and meet the specific needs of the user. By comprehensively obtaining and processing knowledge from multiple information sources, the agent can build a more comprehensive and in-depth user demand model, which not only helps improve the accuracy and satisfaction of flight recommendations but also enhances the intelligence and humanization of aviation services.

[0090] In an alternative embodiment, after extracting knowledge information from the conversation history, business information, user preference information, and status information of each route, the agent can perform semantic splicing processing on at least two pieces of extracted knowledge information according to the semantic content corresponding to the knowledge information to obtain new knowledge information.

[0091] Optionally, semantic content refers to the meaning and context in knowledge information, describing the inherent meaning of the information and its association with the context. In the aviation sales scenario, the semantic content analyzed by the agent may include "user preference for business class", "high requirement for flight punctuality rate", "user's time preference for booking flights", etc. These semantic contents are not limited to the surface keywords, but also involve deeper understanding, such as the potential meaning in the user's words and the implicit needs for services.

[0092] In addition, semantic splicing processing refers to the agent logically combining multiple pieces of knowledge information extracted according to their semantic content to generate more comprehensive and insightful new knowledge information. This process involves the analysis, understanding, and integration of knowledge information, aiming to create a comprehensive knowledge view that can reflect the user's needs and the scenario state. For example, the agent can splice the knowledge information of "user preference for business class" and "flight punctuality rate" to generate the information of "users prefer business class and require a high flight punctuality rate". In this way, the agent can generate new knowledge information that better meets the user's needs and the complexity of the scenario, providing a richer and more accurate basis for subsequent decision-making and recommendations.

[0093] Optionally, the new knowledge information is obtained through the above semantic splicing processing and contains comprehensive information with deeper meanings. The generation of this kind of information not only improves the agent's ability to understand the user's needs but also enables it to provide more personalized services based on a more refined and accurate scenario state. For example, a piece of new knowledge information may be: "Users prefer business class, require a flight punctuality rate higher than 90%, are sensitive to price, and hope to travel on weekdays to avoid peak hours." Such information not only includes the user's preferences for cabin class and punctuality rate but also takes into account price factors and special travel time requirements, which is an important reference for the agent to recommend flights to users.

[0094] In an optional embodiment, filling the knowledge information into a basic template to obtain a target knowledge prompt includes: screening the knowledge information extracted from the conversation history, business information, user preference information, and status information of each flight route according to the business scenario prompt words in the basic template to obtain target content with a correlation coefficient greater than a preset threshold with the business scenario prompt words; filling the target content into the basic template to obtain the target knowledge prompt.

[0095] Optionally, information screening refers to the evaluation and selection of knowledge information extracted from multiple information sources by the agent before generating the target knowledge prompt, ensuring that only content that is highly relevant to the business scenario prompt is retained. This process usually involves the calculation of a correlation coefficient, that is, evaluating the degree of association between the knowledge information and the business scenario prompt. Only when the correlation coefficient of the knowledge information exceeds a preset threshold will they be considered as target content and then filled into the basic template. Target content refers to the knowledge information retained after information screening, which is considered to be highly relevant to the user's intention and business scenario, and is the direct basis for the agent to generate the target knowledge prompt.

[0096] In an optional embodiment, the knowledge information is filled into the basic template to obtain the target knowledge prompt, including: first, the intelligent agent performs information screening on the knowledge information extracted from the business information according to the business scenario prompt in the basic template, and obtains the first content in the knowledge information whose correlation coefficient with the business scenario prompt is greater than the preset threshold. Then, the intelligent agent integrates the business scenario prompt and the first content into the second content according to the basic template, and performs information screening on the knowledge information extracted from the conversation history according to the second content, and obtains the third content in the knowledge information whose correlation coefficient with the second content is greater than the preset threshold. Subsequently, the intelligent agent integrates the second content and the third content into the fourth content according to the basic template, and performs information screening on the knowledge information extracted from the user's preference information and the status information of each route according to the fourth content, and obtains the fifth content in the knowledge information whose correlation coefficient with the fourth content is greater than the preset threshold. Finally, the intelligent agent integrates the fourth content and the fifth content and fills them into the basic template to obtain the target knowledge prompt.

[0097] Optionally, the present application provides an information screening and integration strategy using a "snowball" style, whereby the intelligent agent can achieve knowledge information integration from coarse to fine, from local to overall, significantly improving the accuracy and efficiency of information processing. First, the intelligent agent screens the knowledge information in the business information based on the business scenario prompt words, and the intelligent agent obtains the first content that is most directly related to the user's request. This preliminary screening ensures that the focus of subsequent processing is always on the core needs of the user. Secondly, information screening is not completed in one go, but is carried out in layers. First, the business scenario prompt words are integrated with the first content into the second content to provide more accurate guidance for further session history screening. Next, the second content and the third content are integrated, and then screened and integrated with the user's preference information and route status information to form the fourth content and the fifth content. This process is like a snowball, gradually increasing the richness and pertinence of the information until the target knowledge prompt is finally generated.

[0098] In addition, it should be noted that by hierarchically integrating the user's preference information and flight route status information, the intelligent agent can generate more personalized target knowledge prompts. This prompt not only takes into account the user's business needs but also incorporates the user's personal preferences and real-time environmental information, providing the user with a one-stop and all-round flight query and recommendation service. By adopting a "snowball" screening strategy, the intelligent agent can efficiently screen out the most relevant content from a large amount of information while ensuring the accuracy of the screening process. In each round of screening, the interference of irrelevant information is avoided through the evaluation of the correlation coefficient, ensuring that the finally generated target knowledge prompt is both comprehensive and accurate. This strategy is particularly suitable for dealing with complex and changing business scenarios such as flight query and recommendation. By gradually screening and integrating knowledge information from different sources, the intelligent agent can adapt to the diversity of user needs and consider the dynamic changes of the business scenario, providing users with more flexible and detailed personalized services.

[0099] As can be seen from the above, the "snowball" information screening and integration strategy in this application, through gradual screening and precise integration, ensures that the target knowledge prompt generated by the intelligent agent can accurately reflect the user's needs. At the same time, considering the complexity of the business scenario, it greatly improves the service quality and personalization level of the intelligent agent in the aviation sales field. The application of this strategy not only demonstrates the innovation of the intelligent agent in knowledge management and personalized service but also reflects its efficiency and intelligence in dealing with complex business scenarios.

[0100] In an optional embodiment, the intelligent agent can use a large language model to perform intent parsing and air ticket information reasoning on the target knowledge prompt and air ticket query request according to the first type of prompt words and the second type of prompt words. According to the predicted user intent and air ticket information reasoning result, at least one recommended air ticket information that meets the user intent is determined, where the presentation form of the recommended air ticket information includes at least one of text form, voice form, image form, and video form.

[0101] Optionally, when the presentation form of the recommended flight ticket information is in voice form, the large language model is used to determine the language, tone, and voice style to be used when communicating the recommended flight ticket information to the user based on the target knowledge prompt; when the presentation form of the recommended flight ticket information is in text form, the large language model is used to determine the character form and text expression style to be used when communicating the recommended flight ticket information to the user based on the target knowledge prompt; when the presentation form of the recommended flight ticket information is in image form, the large language model is used to determine the image elements and the combination method of the image elements to be used when communicating the recommended flight ticket information to the user based on the target knowledge prompt; when the presentation form of the recommended flight ticket information is in video form, the large language model is used to determine the audio elements and video elements to be used when communicating the recommended flight ticket information to the user based on the target knowledge prompt.

[0102] Optionally, in the above content, flight ticket information inference means that the intelligent agent uses the large language model to infer the flight ticket information that best meets the user's needs based on the target knowledge prompt and the user's query request, combined with business information, user preferences, and real-time environment information. This inference process is not just a simple data match, but through the powerful inference ability of the model and in-depth understanding of knowledge information, to provide users with more comprehensive and personalized flight ticket recommendations. For example, the large language model may consider the on-time history of flights, special services (such as pet carriage, meal requirements), fare fluctuation trends, etc., to recommend a flight with high comprehensive cost performance for the user.

[0103] Optionally, the presentation form of the recommended flight ticket information refers to the way the intelligent agent communicates the recommended result to the user, including at least one of text form, voice form, image form, and video form. The intelligent agent selects a suitable presentation form according to user preferences and interaction scenarios to enhance the user experience. For example, for users who are used to using voice assistants, the intelligent agent may choose the voice form; for visually oriented users, the image or video form may be used to convey information.

[0104] The intelligent agent can use the large language model to determine a personalized communication method for the recommended flight ticket information in different presentation forms. In the voice form, the intelligent agent predicts the possible language, tone, and voice style preferences of the user based on the target knowledge prompt to ensure that the voice communication is more in line with the user's preferences. In the text form, the intelligent agent may adjust the wording of the text and use the character form and expression style preferred by the user, such as a formal or casual language style. In the image or video form, the intelligent agent will design an image containing specific image elements and combination methods, or select appropriate audio and video elements based on the target knowledge prompt to present the recommended information visually and auditorily.

[0105] When expressing flight ticket recommendation information in voice form, the large language model can predict the language (such as Mandarin or English), tone (professional or friendly), and voice style (male or female voice) that the user may prefer based on the user preferences (such as interest in business class) and business scenarios (such as flight punctuality rate) in the target knowledge prompt. The intelligent agent then uses these prediction results to convey the recommendation results in the voice manner preferred by the user, making the interaction process more user-friendly.

[0106] When expressing flight ticket recommendation information in text form, by analyzing the user preference information in the target knowledge prompt, the large language model can adjust the style of the text expression, such as more formal or more informal wording, to match the user's reading preference. When presenting the recommendation information, the intelligent agent can use the character form preferred by the user (such as Simplified Chinese or English), as well as the text style combined with the user's mood and context, to improve the attractiveness and readability of the text communication.

[0107] When expressing flight ticket recommendation information in image form, for the special needs of the user's query (such as traveling with a pet, requiring wheelchair service), the large language model can design a recommended image containing specific image elements (such as pet logo, wheelchair icon) through the target knowledge prompt. By adjusting the combination method of the image elements, the intelligent agent can intuitively present the features of the recommended flight, such as highlighting business class services or special meal options in the image, enabling the user to obtain key information at a glance.

[0108] When expressing flight ticket recommendation information in video form, the large language model can create recommended content containing audio and video elements based on the target knowledge prompt. The audio elements may include a voice description of the recommended flight, while the video elements can show the internal facilities of the flight, service processes, or introductions of scenic spots at the destination. Through the carefully designed audio and video elements, the intelligent agent can convey the recommended flight ticket information in a more vivid and attractive form, enhancing the user's visual and auditory experience.

[0109] Through the application of the above personalized communication methods, the intelligent agent can not only accurately understand the user's needs and business scenarios through the large language model, but also convey the recommended flight ticket information in the most appropriate presentation form according to the user's preferences and interaction habits, thus significantly improving the personalization level of the service and the user experience. The application of this innovative technology provides a new direction and development space for the intelligent services in the aviation sales field.

[0110] In an alternative embodiment, Figure 3 is a schematic diagram of an alternative intelligent agent working process according to an embodiment of the present application, as Figure 3 shown, the composition of the business expert intelligent agent includes: business scenario prompt words, conversation history (corresponding to Figure 3session records in the database), business information (corresponding to Figure 3 business data generated by the business services in

[0111] Optionally, as Figure 3 shown, the memory module contains a knowledge set component, which stores three key types of knowledge: behavior preference knowledge, business knowledge, and environmental knowledge.

[0112] Business scenario prompt words are written according to the current intelligent agent's business needs, aiming to guide the large model to parse the user's intention and construct a business service request object; at the same time, when processing the business scenario response, they can also guide the large model to perform intelligent reasoning, optimize the response content from multiple perspectives, and ensure its accuracy and effectiveness. The conversation history records all request and response interactions between the user and the system. Business information is relevant business information obtained from external dependent services based on the user's request and the current business context.

[0113] Optionally, the memory module is responsible for constructing the knowledge set by integrating information from multiple original data sources, such as the past conversation interaction records between the user and the business system, the user's business operation history in the business system, and real-time environmental information, etc. Especially environmental information (as Figure 3 shown, it can be obtained from weather services, traffic services, and disaster warning services), which is closely related to factors such as time (e.g., policy restrictions) and geographical location (e.g., local climate, traffic conditions). It can integrate knowledge acquisition capabilities during the knowledge construction stage and perform dynamic information integration during the knowledge application stage. By analyzing this information, personal preference information and taboos of the user, relevant business information such as business habits and behaviors, weather conditions, traffic information, and possible accident warnings are refined, and these contents are transformed into knowledge helpful for specific business scenarios, namely behavior preference knowledge, business knowledge, and environmental knowledge.

[0114] Knowledge application prompt words are written around different classifications and business perspectives of the refined knowledge set to guide the large model to make detailed optimizations in reasoning to fully utilize this knowledge, thereby improving the pertinence and effectiveness of reasoning. For the breadth expansion of the business, knowledge application prompt words need to be written based on different business perspectives, and for the depth expansion of the business, knowledge application prompt words need to be written based on the same business perspective.

[0115] The prompt word synthesizer is mainly used to synthesize prompt words for knowledge and wisdom application when processing the business scenario response. It uses the business scenario prompt words as the basic template, combines multiple knowledge application prompt words from different business perspectives, and is assisted by the user conversation, business information, and knowledge set, and then synthesizes multiple prompt words for knowledge and wisdom application.

[0116] Knowledge and wisdom application prompt words are used to guide the large model to optimize business information from different business perspectives.

[0117] The knowledge set component is responsible for storing the knowledge refined through the memory module and participating in the synthesis of prompt words when the knowledge in the knowledge set is applied.

[0118] An optional service process is as follows:

[0119] Step 1: The business expert agent receives the request information put forward by the user in the session.

[0120] Step 2: Combining the business scenario prompt words and the user session content, the agent requests the large model service to parse the business services to be accessed.

[0121] Step 3: The large model analyzes the business scenario prompt words and the user session content, determines the business services to be accessed and the necessary service request objects, and returns them to the agent in the form of structured data.

[0122] Step 4: The agent uses this structured data to request the corresponding business services.

[0123] Step 5: The business service returns relevant data according to the request content, and the agent processes and stores this data in the business information module.

[0124] Step 6: Through the prompt word synthesizer, the agent integrates the user request, business information, business scenario prompt words, knowledge set content, and knowledge application prompt words to generate multiple knowledge and wisdom application prompt words for different business perspectives.

[0125] Step 7: The business expert agent passes these prompt words to the large model service in multiple rounds for optimizing the user request response. Serial processing is adopted for optimization tasks with dependencies, while parallel processing is adopted for those without dependencies to improve efficiency.

[0126] Step 8: The agent sorts out and combines the user request responses optimized through multiple rounds, and assembles them into the best response finally suitable for returning to the user.

[0127] Step 9: The business application returns the final response to the user, and the user can then initiate a new request.

[0128] As can be seen from the above, through the comprehensive application of business scenario prompt words, conversation history, and business information, the agent in this application can deeply understand the specific needs and preferences of users, rather than just making recommendations based on the origin, destination, and price. For example, users may have habitual information such as a preference for flights of a specific airline, a preference for specific departure or arrival times, and a past special meal requirement, as well as behavioral preference information such as weather preferences and an emphasis on flight punctuality. Through the processing of the large model, the agent can accurately capture and apply these preferences, significantly improving the personalization level of ticket recommendations. Secondly, in this application, the agent not only focuses on user preferences but also obtains the status information of each route in real time, including but not limited to flight punctuality, aircraft type information, cabin status, fare fluctuations, and early warnings of emergencies. This means that when recommending tickets, the system can comprehensively consider various factors, provide more comprehensive and accurate flight information, and help users make more reasonable itinerary choices.

[0129] According to another aspect of the present application, there is also provided a ticket recommendation device based on a large language model, wherein, Figure 4 is a schematic diagram of an optional ticket recommendation device based on a large language model according to an embodiment of the present application, as Figure 4 shown, including: a first acquisition unit 401, a second acquisition unit 402, a first processing unit 403, and a second processing unit 404.

[0130] Optionally, the first acquisition unit 401 is configured to, in response to a ticket query request initiated by a user, obtain business scenario prompt words, conversation history, and business information through the agent, where the business scenario prompt words are used to represent the business scenario information corresponding to the agent, the conversation history is the historical interaction information between the user and the software system, and the business information is information related to the aviation business obtained from application services other than the agent according to the ticket query request and the business scenario information corresponding to the agent; the second acquisition unit 402 is configured to obtain the preference information of the user and the status information of each route, where the preference information at least includes habitual information of the user when handling aviation business and behavioral preference information of the user when choosing air travel; the first processing unit 403 is configured to generate a basic template based on the business scenario prompt words, extract knowledge information from the conversation history, business information, preference information of the user, and the status information of each route, and fill the knowledge information into the basic template to obtain a target knowledge prompt; the second processing unit 404 is configured to process the ticket query request through the large language model based on the target knowledge prompt to generate at least one recommended ticket information for the user.

[0131] Optionally, the second acquisition unit 402 includes: a first determination subunit, configured to determine, by means of keyword recognition, the user's habitual information when handling air business and the user's behavioral preference information when selecting air travel from the conversation history and the user's historical itinerary information, so as to obtain the user's preference information; a second determination subunit, configured to determine the ticket price information, flight on-time rate, flight delay risk information, and weather information along the route for each route from the business information; a third determination subunit, configured to use the ticket price information, flight on-time rate, flight delay risk information, and weather information along the route for each route as the status information of each route.

[0132] Optionally, the first processing unit 403 includes: a first acquisition subunit, configured to acquire a first type of prompt word, a second type of prompt word, and a third type of prompt word, where the first type of prompt word is used to prompt the parsing method adopted by the large language model when parsing the user's intention, the second type of prompt word is used to prompt the inference angle selected by the large language model when processing the target knowledge prompt, and the third type of prompt word is used to constrain the data format filled into the basic template; a first generation subunit, configured to generate a basic template according to the first type of prompt word, the second type of prompt word, the third type of prompt word, and the business scenario prompt word.

[0133] Optionally, the first processing unit 403 is further configured to extract corresponding knowledge information from each of the conversation history, business information, user preference information, and status information of each route using the same air business perspective; and / or extract corresponding knowledge information from each of the conversation history, business information, user preference information, and status information of each route using different air business perspectives.

[0134] Optionally, the air ticket recommendation device based on the large language model further includes: a semantic splicing unit, configured to perform semantic splicing processing on at least two pieces of extracted knowledge information according to the semantic content corresponding to the knowledge information, so as to obtain new knowledge information.

[0135] Optionally, the first processing unit 403 includes: a first processing subunit, configured to perform information screening on the knowledge information extracted from the conversation history, business information, user preference information, and status information of each route according to the business scenario prompt word in the basic template, so as to obtain target content with a correlation coefficient greater than a preset threshold with the business scenario prompt word; a second processing subunit, configured to fill the target content into the basic template to obtain a target knowledge prompt.

[0136] Optionally, the first processing unit 403 includes: a third processing subunit, configured to perform information screening on the knowledge information extracted from the service information according to the service scenario prompt words in the basic template, and obtain the first content in the knowledge information whose correlation coefficient with the service scenario prompt words is greater than a preset threshold; a fourth processing subunit, configured to integrate the service scenario prompt words and the first content into the second content according to the basic template, and perform information screening on the knowledge information extracted from the conversation history according to the second content, and obtain the third content in the knowledge information whose correlation coefficient with the second content is greater than a preset threshold; a fifth processing subunit, configured to integrate the second content and the third content into the fourth content according to the basic template, and perform information screening on the knowledge information extracted from the user's preference information and the status information of each flight route according to the fourth content, and obtain the fifth content in the knowledge information whose correlation coefficient with the fourth content is greater than a preset threshold; a sixth processing subunit, configured to integrate the fourth content and the fifth content and fill them into the basic template to obtain the target knowledge prompt.

[0137] Optionally, the second processing unit 404 includes: a fourth determination subunit, configured to perform intent parsing and flight ticket information reasoning on the target knowledge prompt and the flight ticket query request through a large language model according to the first type of prompt words and the second type of prompt words, and determine at least one recommended flight ticket information that meets the user's intent according to the predicted user intent and the flight ticket information reasoning result, where the presentation form of the recommended flight ticket information includes at least one of text form, voice form, image form, and video form.

[0138] Optionally, the flight ticket recommendation device based on a large language model further includes: a first determination unit, configured to, when the presentation form of the recommended flight ticket information is in voice form, determine the language, tone, and voice style used when conveying the recommended flight ticket information to the user through voice according to the target knowledge prompt through the large language model; a second determination unit, configured to, when the presentation form of the recommended flight ticket information is in text form, determine the character form and text expression style used when conveying the recommended flight ticket information to the user through text according to the target knowledge prompt through the large language model; a third determination unit, configured to, when the presentation form of the recommended flight ticket information is in image form, determine the image elements and the combination manner of the image elements used when conveying the recommended flight ticket information to the user through image according to the target knowledge prompt through the large language model; a fourth determination unit, configured to, when the presentation form of the recommended flight ticket information is in video form, determine the audio elements and video elements used when conveying the recommended flight ticket information to the user through video according to the target knowledge prompt through the large language model.

[0139] According to another aspect of the present application, there is also provided a computer-readable storage medium, wherein the computer-readable storage medium includes a stored executable program, and when the executable program runs, it controls the device where the computer-readable storage medium is located to execute the above-mentioned air ticket recommendation method based on a large language model.

[0140] According to another aspect of the present application, there is also provided an electronic device, including: a memory storing an executable program; a processor for running the program, and when the program runs, it executes the above-mentioned air ticket recommendation method based on a large language model.

[0141] According to another aspect of the present application, there is also provided a computer program product, including computer instructions, characterized in that when the computer instructions are executed by a processor, the steps of the above-mentioned air ticket recommendation method based on a large language model are implemented.

[0142] The serial numbers of the above-mentioned embodiments of the present application are only for description and do not represent the advantages or disadvantages of the embodiments.

[0143] In the above-mentioned embodiments of the present application, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0144] In the several embodiments provided by the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only illustrative. For example, the division of the units can be a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of the units or modules can be in electrical or other forms.

[0145] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place, or can be distributed to multiple units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0146] In addition, the functional units in the various embodiments of the present application can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0147] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), mobile hard disks, magnetic disks, or optical discs.

[0148] The above are only the preferred embodiments of this application. It should be noted that for those of ordinary skill in the art of this technology, without departing from the principle of this application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of this application.

Claims

1. A flight ticket recommendation method based on large language models, characterized in that, Including: In response to a flight ticket query request initiated by a user, obtain a business scenario prompt word, conversation history, and business information through an agent. Among them, the business scenario prompt word is used to represent the business scenario information corresponding to the agent, the conversation history is the historical interaction information between the user and the software system, and the business information is information related to the aviation business obtained from application services other than the agent according to the flight ticket query request and the business scenario information corresponding to the agent; Obtain the user's preference information and the status information of each flight route. Among them, the preference information at least includes the user's habitual information when handling aviation business and the user's behavioral preference information when selecting air travel; Generate a basic template based on the business scenario prompt word, extract knowledge information from the conversation history, the business information, the user's preference information, and the status information of each flight route, and fill the knowledge information into the basic template to obtain a target knowledge prompt; Process the flight ticket query request through a large language model based on the target knowledge prompt to generate at least one recommended flight ticket information for the user; Among them, generating a basic template based on the business scenario prompt word includes: obtaining a first type of prompt word, a second type of prompt word, and a third type of prompt word. Among them, the first type of prompt word is used to prompt the parsing method adopted by the large language model when parsing the user's intention, the second type of prompt word is used to prompt the inference angle selected by the large language model when processing the target knowledge prompt, and the third type of prompt word is used to constrain the data format filled into the basic template; generate the basic template according to the first type of prompt word, the second type of prompt word, the third type of prompt word, and the business scenario prompt word; Among them, extracting knowledge information from the conversation history, the business information, the user's preference information, and the status information of each flight route includes: extracting corresponding knowledge information from each of the conversation history, the business information, the user's preference information, and the status information of each flight route using the same aviation business perspective; and / or, extracting corresponding knowledge information from each of the conversation history, the business information, the user's preference information, and the status information of each flight route using different aviation business perspectives; Fill the knowledge information into the basic template to obtain the target knowledge prompt, including: screening the knowledge information extracted from the business information according to the business scenario prompt words in the basic template to obtain the first content in the knowledge information whose correlation coefficient with the business scenario prompt words is greater than a preset threshold; integrating the business scenario prompt words and the first content according to the basic template to obtain the second content, and screening the knowledge information extracted from the conversation history according to the second content to obtain the third content in the knowledge information whose correlation coefficient with the second content is greater than the preset threshold; integrating the second content and the third content according to the basic template to obtain the fourth content, and screening the knowledge information extracted from the user's preference information and the status information of each flight route according to the fourth content to obtain the fifth content in the knowledge information whose correlation coefficient with the fourth content is greater than the preset threshold; integrating the fourth content and the fifth content and filling them into the basic template to obtain the target knowledge prompt.

2. The air ticket recommendation method based on a large language model according to claim 1, wherein Obtain the user's preference information and the status information of each flight route, including: Determine the user's habitual information when handling air business and the user's behavioral preference information when choosing air travel from the conversation history and the user's historical itinerary information through keyword recognition to obtain the user's preference information; Determine the ticket price information, flight punctuality rate, flight delay risk information, and weather information along the route of each flight route from the business information; Use the ticket price information, flight punctuality rate, flight delay risk information, and weather information of each flight route as the status information of each flight route.

3. The air ticket recommendation method based on a large language model according to claim 1, wherein After extracting knowledge information from the conversation history, the business information, the user's preference information, and the status information of each flight route, the air ticket recommendation method based on the large language model further includes: Perform semantic splicing processing on at least two pieces of the extracted knowledge information according to the semantic content corresponding to the knowledge information to obtain new knowledge information.

4. The air ticket recommendation method based on a large language model according to claim 1, wherein Fill the knowledge information into the basic template to obtain the target knowledge prompt, including: Screen the knowledge information extracted from the conversation history, the business information, the user's preference information, and the status information of each flight route according to the business scenario prompt words in the basic template to obtain the target content whose correlation coefficient with the business scenario prompt words is greater than a preset threshold; Fill the target content into the basic template to obtain the target knowledge prompt.

5. The air ticket recommendation method based on a large language model according to claim 1, wherein Process the air ticket query request by the large language model according to the target knowledge prompt to generate at least one recommended air ticket information for the user, including: Based on the large language model, according to the first type of prompt words and the second type of prompt words, perform intent parsing and air ticket information reasoning on the target knowledge prompt words and the air ticket query request, and determine at least one recommended air ticket information that meets the user's intent. The presentation forms of the recommended air ticket information include at least one of text form, voice form, image form, and video form.

6. The air ticket recommendation method based on a large language model according to claim 5, wherein The air ticket recommendation method based on the large language model further includes: When the presentation form of the recommended air ticket information is in voice form, based on the target knowledge prompt words, determine the language, tone, and voice style to be used when conveying the recommended air ticket information to the user by voice through the large language model; When the presentation form of the recommended air ticket information is in text form, based on the target knowledge prompt words, determine the character form and text expression style to be used when conveying the recommended air ticket information to the user by text through the large language model; When the presentation form of the recommended air ticket information is in image form, based on the target knowledge prompt words, determine the image elements and the combination method of the image elements to be used when conveying the recommended air ticket information to the user by image through the large language model; When the presentation form of the recommended air ticket information is in video form, based on the target knowledge prompt words, determine the audio elements and video elements to be used when conveying the recommended air ticket information to the user by video through the large language model.

7. An air ticket recommendation device based on a large language model, characterized in that, It includes: A first acquisition unit, configured to, in response to an air ticket query request initiated by a user, obtain business scenario prompt words, conversation history, and business information through an intelligent agent, where the business scenario prompt words are used to represent the business scenario information corresponding to the intelligent agent, the conversation history is the historical interaction information between the user and the software system, and the business information is information related to the aviation business obtained from application services other than the intelligent agent according to the air ticket query request and the business scenario information corresponding to the intelligent agent; A second acquisition unit, configured to obtain the user's preference information and the status information of each flight route, where the preference information at least includes the user's habit information when handling aviation business and the user's behavioral preference information when choosing air travel; A first processing unit, configured to generate a basic template based on the business scenario prompt words, extract knowledge information from the conversation history, the business information, the user's preference information, and the status information of each flight route, and fill the knowledge information into the basic template to obtain a target knowledge prompt word; A second processing unit, configured to process the air ticket query request through the large language model based on the target knowledge prompt word to generate at least one recommended air ticket information for the user; Among them, the first processing unit includes: a first acquisition subunit, configured to acquire a first type of prompt word, a second type of prompt word, and a third type of prompt word, where the first type of prompt word is used to prompt the large language model to adopt a parsing method when parsing the user's intention, the second type of prompt word is used to prompt the large language model to select an inference angle when processing the target knowledge prompt, and the third type of prompt word is used to constrain the data format filled into the basic template; a first generation subunit, configured to generate a basic template according to the first type of prompt word, the second type of prompt word, the third type of prompt word, and the business scenario prompt word; The first processing unit is further configured to extract corresponding knowledge information from each of the session history, business information, user preference information, and status information of each flight route using the same aviation business perspective; and / or extract corresponding knowledge information from each of the session history, business information, user preference information, and status information of each flight route using different aviation business perspectives; Among them, the first processing unit includes: a third processing subunit, configured to screen the knowledge information extracted from the business information according to the business scenario prompt word in the basic template to obtain a first content in the knowledge information whose correlation coefficient with the business scenario prompt word is greater than a preset threshold; a fourth processing subunit, configured to integrate the business scenario prompt word and the first content into a second content according to the basic template, and screen the knowledge information extracted from the session history according to the second content to obtain a third content in the knowledge information whose correlation coefficient with the second content is greater than the preset threshold; a fifth processing subunit, configured to integrate the second content and the third content into a fourth content according to the basic template, and screen the knowledge information extracted from the user preference information and the status information of each flight route according to the fourth content to obtain a fifth content in the knowledge information whose correlation coefficient with the fourth content is greater than the preset threshold; a sixth processing subunit, configured to integrate the fourth content and the fifth content and fill them into the basic template to obtain the target knowledge prompt.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, wherein when the executable program runs, it controls the device where the computer-readable storage medium is located to execute the large language model-based air ticket recommendation method according to any one of claims 1 to 6.

9. An electronic device, characterized in that, Including: A memory storing an executable program; A processor for running the program, wherein when the program runs, it executes the large language model-based air ticket recommendation method according to any one of claims 1 to 6.

10. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, the steps of the large language model-based air ticket recommendation method according to any one of claims 1 to 6 are implemented.

Images