Intelligent travel service data processing method, system and equipment

Abstract

The invention provides an intelligent travel service data processing method, system and device, and belongs to the technical field of data processing. The method comprises the following steps: receiving multi-mode travel request data of voice, images and texts, and converting the data into unified semantic representation data through feature extraction and fusion; deep analysis is carried out to construct a structured demand semantic model; complementing information through multiple rounds of conversations to obtain complete service request data; then aggregating with multi-source service data, generating an optimal scheme through multi-dimensional constraint optimization, dynamically presenting through interactive processing, and performing optimization adjustment according to user feedback; then realizing service execution and generating execution data; and finally, finishing travel reimbursement business based on an event triggering mechanism. According to the invention, intelligent and efficient travel service data processing is realized.

Description

Technical Field

[0001] This invention belongs to the field of data processing technology, and more specifically relates to a smart travel service data processing method, system and equipment. Background Technology

[0002] As enterprises deepen their digital transformation, various integrated service platforms have emerged in the travel management field. These platforms aggregate supply chain resources such as air tickets and hotels, providing corporate users with a unified booking portal and playing a positive role in centralized resource procurement and cost control.

[0003] However, existing platforms still have several limitations in their technical implementation. Their interaction methods are primarily based on traditional graphical user interfaces, requiring users to complete operations through standardized controls such as form filling and option selection. This model results in a lengthy operation process, requiring users to navigate through multiple page jumps and repetitive input, making it unsuitable for quickly responding to the travel needs expressed in natural language.

[0004] Regarding rule enforcement, existing systems typically employ a post-event notification mechanism. When a user selects an option that does not comply with company regulations, the system will issue a warning, but it cannot proactively filter and avoid such violations during the solution generation phase. This approach somewhat impacts the effectiveness of rule enforcement.

[0005] Furthermore, the existing system falls short in providing personalized services. While it can store basic, frequently used information, it struggles to make comprehensive decisions based on contextual information, historical behavioral data, and real-time needs. The system typically returns a raw list of data that matches the query criteria, rather than an optimized solution that has undergone intelligent filtering and sorting.

[0006] It's worth noting that current advanced technical solutions employ basic question-answering robots or command-line query functions. While these solutions can recognize simple, fixed instructions, their ability to understand complex natural language statements is limited, and they only support single-turn or limited-turn interactions. When the information provided by the user is incomplete, the system cannot proactively retrieve the missing information through multiple rounds of dialogue, leading to service interruptions. Furthermore, these solutions have limited integration depth with backend business systems, making it difficult to achieve end-to-end automated processing from requirement understanding to service execution. Summary of the Invention

[0007] To address the above issues, the present invention aims to provide an intelligent travel service data processing method, system, and device. Through multimodal data fusion, deep semantic parsing, multi-turn dialogue management, multi-objective optimization decision-making, and smart contracts, the invention achieves intelligent and automated operation of the entire travel service process from demand understanding to reimbursement closed loop, significantly improving service efficiency, compliance, and user experience.

[0008] To achieve the above objectives, the present invention employs the following technical solution: In a first aspect, embodiments of this application provide a smart travel service data processing method, including: It receives multimodal travel request data containing voice, images, and text, and converts it into unified semantic representation data through feature extraction and fusion processing; Deeply analyze and process semantic representation data, and construct structured demand semantic model data through semantic understanding and relationship mining; Multi-turn dialogue processing is performed based on semantic model data, and complete service request data is obtained through information completion and verification. The complete service request data is aggregated with multi-source service data, and the optimal solution data is generated through multi-dimensional constraint optimization. The optimal solution data is dynamically presented through interactive processing, and optimized and adjusted based on user feedback to generate solution data to be confirmed. The system collaboratively processes the final solution data confirmed by the user to execute the service and generate service execution data. Based on the event-triggered mechanism, the service execution data is processed post-event to complete the travel expense reimbursement business.

[0009] In an optional implementation, multimodal travel request data containing voice, images, and text is received and converted into unified semantic representation data through feature extraction and fusion processing. Deeply analyze and process semantic representation data, and construct structured demand semantic model data through semantic understanding and relationship mining; Multi-turn dialogue processing is performed based on semantic model data, and complete service request data is obtained through information completion and verification. The complete service request data is aggregated with multi-source service data, and the optimal solution data is generated through multi-dimensional constraint optimization. The optimal solution data is dynamically presented through interactive processing, and optimized and adjusted based on user feedback to generate solution data to be confirmed. The system collaboratively processes the final solution data confirmed by the user to execute the service and generate service execution data. Based on the event-triggered mechanism, the service execution data is processed post-event to complete the travel expense reimbursement business.

[0010] In an optional implementation, the deep parsing processing of the semantic representation data to construct a structured demand semantic model data through semantic understanding and relation mining includes: A Transformer-based multi-task deep learning model is used to process semantic data packets in parallel, obtaining intent classification data, entity sequence data, and sentiment feature data, among which: The intent classification data is used to guide the focus and scope of subsequent named entity recognition processing; The entity sequence data serves as the basic input data for constructing the semantic relationship graph; The sentiment feature data is used to assist in optimizing the weight configuration of semantic associations; Entity data is extracted from the entity sequence data using named entity recognition technology, and semantic relationship graph data between entities is established by combining the focus defined by the intent classification data. Using graph neural networks to perform multi-hop reasoning on the semantic relationship graph data, and with the assistance of the sentiment feature data, construct demand semantic model data with accurate semantic association; Intermediate and final result data generated during data processing are stored in a structured manner to form a unified and queryable semantic graph dataset.

[0011] In an optional implementation, the multi-turn dialogue processing based on semantic model data, obtaining complete service request data through information completion and verification, includes: The dialogue state is initialized based on the semantic graph dataset, and dialogue state vector data containing the currently identified entities, intents and context is generated, which is denoted as the current state data. The current state data is checked for completeness using a slot verification algorithm. The identified entities are matched and compared with the slots of the required information to complete the current intent, and missing information identification data is generated. The dialogue management model based on reinforcement learning strategy generates optimal follow-up questioning scripts based on the missing information identification data and the current state data, and initiates inquiries to the user through a natural language interaction interface. The system receives user feedback data in real time, updates the dialogue state vector data using an incremental learning algorithm, and dynamically adjusts the follow-up questioning strategy. The process of iteratively completing and verifying information continues until all required information slots are correctly filled, ultimately outputting a complete set of service request data.

[0012] In an optional implementation, the step of aggregating complete service request data with multi-source service data and generating optimal solution data through multi-dimensional constraint optimization includes: The complete set of service request data is aggregated with real-time service data from multiple suppliers obtained through a distributed data acquisition interface to generate aggregated service data. The real-time service data includes flight data from multiple airline ticket suppliers, room availability and price data from multiple hotel suppliers, and vehicle dispatch data from multiple car service suppliers. The aggregated service data will be compared with the cabin class constraint data, accommodation standard constraint data, and transportation standard constraint data in the enterprise travel rules database for compliance. The aggregated service data is filtered in real time using a rule engine to remove data that does not meet the rule constraints and generate a candidate solution dataset. Based on historical travel records, personal preference data, and frequently used cost center data in the user profile database, a federated learning mechanism is used to perform feature collaborative calculation on the aggregated service data. Based on the calculated features, a multi-objective optimization algorithm is used to perform weighted scoring and ranking of the candidate solution dataset in terms of price, time, policy compliance, and personalized preference dimensions. The final result is a queue of optimal solutions that satisfy multiple constraints, containing 1-3 recommended solutions and their parameter data.

[0013] In an optional implementation, the step of dynamically presenting the optimal solution data through interactive processing and optimizing and adjusting it based on user feedback to generate solution data to be confirmed includes: The optimal solution queue data is transformed into a visual, structured display data using an adaptive rendering engine; Receive real-time feedback data streams generated by users through natural language or interactive operations; Based on user feedback data streams, the scheme weight parameters are dynamically adjusted through incremental learning algorithms; A deep learning ranking model is used to optimize and adjust the display order of the optimal solution data in real time, forming the final solution data confirmed by the user. Record user behavior data throughout the interaction process to optimize subsequent recommendation model parameters.

[0014] In an optional implementation, the step of executing the service by processing the user-confirmed final solution data through system collaboration and generating service execution data includes: The data conversion engine standardizes the final solution data confirmed by the user into the message format data required by the target system interface. A secure data transmission channel is established using a two-way authentication mechanism to ensure the security of data transmission; By using blockchain technology to record key operational data nodes during the service execution process, immutable execution evidence data is generated. Utilize workflow engines to drive automated synchronization of business data across multiple enterprise systems.

[0015] In an optional implementation, the step of completing the travel expense reimbursement process by processing service execution data after the fact, based on the event-triggered mechanism, includes: Monitor service status in real time using event listeners and capture service status change event data; Based on service status change event data, relevant trip and expense data are extracted from the business database, and a structured reimbursement dataset is generated after data cleaning and transformation. Smart contract technology is used to automate compliance verification of expense reimbursement data; The verified reimbursement data is pushed to the target system through the system interface to complete the business closed-loop processing.

[0016] Secondly, embodiments of this application also provide an intelligent travel service data processing system, including: The multimodal data receiving and processing module is used to receive multimodal travel request data containing voice, images and text, and convert it into unified semantic representation data through feature extraction and fusion processing; The deep semantic parsing module is used to perform deep parsing processing on semantic representation data, and to build structured demand semantic model data through semantic understanding and relationship mining; The multi-turn dialogue management module is used to process multi-turn dialogues based on semantic model data, and obtain complete service request data through information completion and verification. The multi-source data aggregation and solution generation module is used to aggregate complete service request data with multi-source service data and generate optimal solution data through multi-dimensional constraint optimization. The interactive solution presentation module is used to dynamically present the optimal solution data through interactive processing, and to optimize and adjust it based on user feedback, generating solution data to be confirmed. The service execution and collaboration module is used to execute the final solution data confirmed by the user through system collaboration and generate service execution data. The post-event data processing module is used to complete the travel expense reimbursement business by processing the service execution data after the event triggering mechanism.

[0017] Thirdly, embodiments of this application also provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the intelligent travel service data processing method described in any of the above descriptions.

[0018] Fourthly, embodiments of this application also provide a storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the intelligent travel service data processing method as described in any of the above claims.

[0019] As can be seen from the above technical solutions, the present invention has the following advantages: The intelligent travel service data processing method provided in this application achieves accurate understanding of users' natural language requests through multimodal data fusion and deep semantic parsing; it automatically completes the entire process from requirement completion and multi-source solution aggregation and optimization to service execution through multi-turn dialogue and intelligent decision-making; and finally, through system collaboration and event triggering mechanisms, it realizes automated closed-loop processing of travel reimbursement business, significantly improving the efficiency, compliance and user experience of travel management.

[0020] This application utilizes multimodal data fusion and deep semantic parsing technology to accurately understand the complex travel needs expressed by users in natural language, voice, or images, avoiding the cumbersome operations of traditional form-based interactions and significantly improving interaction efficiency and user experience.

[0021] This application employs a compliance verification and multi-objective optimization algorithm based on a rule engine, which can proactively filter out options that do not comply with corporate regulations from the source before the plan is generated, ensuring that the travel plan automatically complies with preset rules, thereby improving the standardization and reliability of management.

[0022] This application combines user profiles, historical behavioral data, and real-time contextual information, and utilizes federated learning and deep learning ranking models to provide users with dynamically optimized personalized recommendation solutions, significantly improving the intelligence level of services and user satisfaction.

[0023] This application has achieved end-to-end automated processing, from understanding requirements, completing multi-turn dialogues, generating and optimizing solutions, to service execution and post-event reimbursement processing. The entire process has achieved a high degree of automation and system collaboration, reducing manual intervention and significantly improving business processing efficiency and overall collaboration.

[0024] This application uses blockchain technology to generate immutable service execution records and utilizes smart contracts to automatically verify the compliance of business data, ensuring the traceability, transparency, and security of the data processing process, and enhancing the system's credibility and risk control capabilities. Attached Figure Description

[0025] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 A flowchart illustrating the intelligent travel service data processing method provided in this application.

[0027] Figure 2 A schematic diagram of the structure of the intelligent travel service data processing system provided in this application.

[0028] Figure 3 A schematic diagram of the structure of the electronic device provided in this application. Detailed Implementation

[0029] The various embodiments of this disclosure will be described more fully in the detailed steps of the intelligent travel service data processing method described below. This disclosure may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of this disclosure to the specific embodiments disclosed herein, but rather this disclosure should be understood to cover all adjustments, equivalents, and / or alternatives falling within the spirit and scope of the various embodiments of this disclosure.

[0030] In the following, the terms “comprising” or “may include”, which may be used in various embodiments of this disclosure, indicate the presence of the disclosed functions, operations, or elements, and do not limit the addition of one or more functions, operations, or elements. Furthermore, as used in various embodiments of this disclosure, the terms “comprising,” “having,” and their cognates are intended only to indicate a particular feature, number, step, operation, element, component, or combination of the foregoing, and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, elements, components, or combinations of the foregoing, or the possibility of adding one or more combinations of the foregoing.

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] Please see Figure 1 The diagram shown is a flowchart of a smart travel service data processing method according to a specific embodiment. The method includes: S1: Receives multimodal travel request data containing voice, images, and text, and converts it into unified semantic representation data through feature extraction and fusion processing.

[0033] In specific implementations, multimodal travel requests containing voice, images, and text can be sent through front-end entry points integrated into enterprise chat tools (such as DingTalk and WeChat Work), standalone apps, or web pages.

[0034] For the voice input part, an end-to-end speech recognition model is used to first extract acoustic features from the speech, that is, analyze elements such as frequency, pitch, and speech rate, and then convert it into text. For example, if a user says "I plan to book a flight to Shenzhen next Monday", the model can convert this speech into corresponding text content.

[0035] For image input, optical character recognition (OCR) technology is employed. First, preprocessing operations such as grayscale conversion and noise reduction are performed on the image to improve image quality and facilitate subsequent recognition. Next, image feature extraction and text recognition are conducted to convert the text information in the image into structured text. For example, if a user uploads an image containing a travel itinerary, the system can identify key information such as the date and location and organize it into a structured format.

[0036] Next, an attention mechanism is used to weight and fuse features from the multimodal data. Weights are assigned to each feature based on the importance of the information contained in different modalities. For example, if the voice emphasizes an urgent travel need, then voice features related to urgency will have higher weights. This generates unified semantic representation vector data, which is then standardized to fit a specific data distribution range and encapsulated to form semantic data packages containing conversational context information, such as recording the user's previous booking history on the platform, providing richer background information for subsequent semantic understanding.

[0037] S2: Perform deep parsing processing on semantic representation data, and construct structured demand semantic model data through semantic understanding and relationship mining.

[0038] In this specific implementation, a Transformer-based multi-task deep learning model is used to process semantic data packets in parallel. This model can simultaneously output intent classification data, entity sequence data, and sentiment feature data. Intent classification data clearly identifies the user's core purpose, such as booking a flight or searching for a hotel, which determines the focus of subsequent named entity recognition. For example, if the intent is to book a hotel, subsequent recognition will focus on information related to hotel booking.

[0039] Entity sequence data encompasses various entities identified from text, such as departure point, destination, and date. These entities form the basis for constructing semantic relationship graphs. Sentiment feature data reflects users' emotional tendencies towards the service; if users express dissatisfaction, subsequent processing strategies will be adjusted accordingly.

[0040] Next, named entity recognition technology is used to extract entities from entity sequence data, and semantic relationship graphs between entities are constructed by combining the key points determined by intent classification. For example, for the sentence "Book a flight from Chengdu to Hangzhou, departing on the 15th of this month", entities such as "Chengdu", "Hangzhou" and "the 15th of this month" are identified, and semantic relationships of departure location-destination-departure time are constructed based on the intent of "booking a flight".

[0041] Subsequently, graph neural networks are used to perform multi-hop inference on the semantic relationship graph data to uncover deeper semantic connections. With the assistance of sentiment feature data, a demand semantic model data with accurate semantic connections is constructed. For example, if sentiment features indicate that users are more sensitive to price, the model will focus more on price-related semantic connections during its construction.

[0042] The intermediate and final results generated throughout the process will be stored in a structured manner to form a unified and queryable semantic graph dataset, so that relevant information can be retrieved in subsequent steps.

[0043] S3: Performs multi-turn dialogue processing based on semantic model data, and obtains complete service request data through information completion and verification.

[0044] In a specific implementation, the dialogue state is initialized based on the semantic graph dataset, generating dialogue state data that includes identified entities, intents, and context. For example, if the identified intent is to book a hotel and the destination is identified as "Nanjing," the dialogue state data will include this information and relevant background.

[0045] Then, using a slot verification algorithm, the identified entities are matched against the required information slots needed to complete the current intent, thereby checking the completeness of the dialogue information. For hotel bookings, required information slots might include destination, check-in date, check-out date, hotel star rating, etc. If only the destination is identified, a missing information flag will be generated.

[0046] A dialogue management model based on reinforcement learning generates optimal follow-up questions based on missing information identifiers and the current dialogue state, and then asks the user through natural language interaction. For example, if check-in date information is missing, the model might generate a follow-up question like, "When do you plan to check into the hotel?"

[0047] The system receives user feedback in real time, updates dialogue state data using incremental learning algorithms, and dynamically adjusts follow-up question strategies. For example, after a user replies with their check-in date, the algorithm integrates this information into the dialogue state and reassesses whether there is any other missing information, thereby determining the content of subsequent follow-up questions.

[0048] The above information completion and verification process is continuously iterated until all required information slots are correctly filled, and finally a complete set of service request data is output, covering all the user's travel service requirements.

[0049] S4: Aggregate complete service request data with multi-source service data, and generate optimal solution data through multi-dimensional constraint optimization.

[0050] In a specific implementation, complete service request data is aggregated with real-time service data from multiple suppliers acquired through distributed data collection. Distributed data collection simultaneously connects to multiple airline, hotel, and car rental suppliers to obtain flight data, room availability and pricing data, vehicle dispatch data, etc. For example, it acquires flight information from Beijing to Guangzhou from multiple airline suppliers for a user-specified date, including flight time, price, and cabin class; and room availability and pricing from multiple hotel suppliers for the same time period.

[0051] After aggregation, this service data is compared for compliance with constraints such as cabin class, accommodation standards, and transportation standards in the company's travel rules database. For example, it checks whether the booked hotel price exceeds the company's accommodation standards and whether the flight cabin class meets the employee's level restrictions.

[0052] The aggregated service data is filtered in real time using a rules engine to remove data that does not meet the rules and generate a candidate solution dataset. For example, hotel options that exceed the accommodation standards will be removed from the candidate solutions.

[0053] Next, based on historical travel records, personal preferences, and frequently used cost center data in the user profile database, a federated learning mechanism is used to perform feature collaborative computation on the aggregated service data. For example, by combining users' historical flight booking cabin class preferences and budget constraints of frequently used cost centers, feature extraction and collaborative analysis are performed on the service data.

[0054] Based on the calculated features, a multi-objective optimization algorithm is used to weight and rank the candidate solution dataset across dimensions such as price, time, policy compliance, and personalized preferences. Weights are assigned according to the importance of each dimension to the user, and each candidate solution is comprehensively evaluated. For example, for users who prioritize price, the weight of the price dimension is increased; for users who are time-sensitive, the weight of the time dimension is increased.

[0055] The final result is a queue of optimal solutions that meet multiple constraints. It includes 1-3 recommended solutions and their detailed parameters, such as flight information, hotel details, and transportation arrangements, providing users with the best choice that takes into account corporate policies, personal preferences, and real-time service availability.

[0056] S5: Dynamically present the optimal solution data through interactive processing, and optimize and adjust it based on user feedback to generate solution data to be confirmed.

[0057] In a specific implementation, the optimal solution queue data is converted into visualized structured display data through an adaptive rendering engine. Depending on different terminal devices (such as mobile phones and computers) and display environments, it is presented in an appropriate format. For example, on mobile devices, it is displayed in the form of cards, and on computers, it is displayed in the form of lists, making it convenient for users to view and compare information such as flight times, hotel names and prices.

[0058] It receives real-time feedback from users through natural language or interactive actions. For example, a user might use voice to say, "I want a more cost-effective solution," or click to view detailed information about a solution.

[0059] Based on user feedback, the algorithm dynamically adjusts the weight parameters of the proposed solutions using an incremental learning algorithm. If a user desires a more cost-effective solution, the algorithm will adjust the weights of dimensions such as price and service quality accordingly, and recalculate the solution score.

[0060] Then, a deep learning ranking model is used to optimize and adjust the display order of the optimal solution data in real time. Based on the adjusted weight parameters and solution characteristics, the solutions that better meet the user's needs are ranked first. For example, after weight adjustment, the solution with better cost performance is displayed first, forming the final solution data confirmed by the user.

[0061] Throughout the interaction, user behavior data is recorded, such as the time spent viewing options, clicks, and feedback. This data is used to optimize subsequent recommendation model parameters, improving the accuracy and personalization of recommendations. For example, if a user frequently chooses a certain type of hotel, subsequent recommendations will appropriately increase the recommendation weight for that type of hotel.

[0062] S6: The system will process the final solution data confirmed by the user to execute the service and generate service execution data.

[0063] In a specific implementation, a data conversion engine standardizes the final plan data confirmed by the user into a message format required by the target system interface. Different backend business systems (such as booking engines and expense reimbursement systems) have different data interface requirements, and the data conversion engine performs format conversion according to the target system specifications. For example, it converts booking plan data into JSON format data that the booking engine API can accept.

[0064] A secure data transmission channel is established using a two-way authentication mechanism to ensure data security. Two-way authentication requires both communicating parties to verify each other's identities to prevent data theft or tampering. For example, when using the SSL / TLS protocol for encrypted communication, the client and server verify each other's certificates before transmitting data.

[0065] Blockchain technology records key operational data nodes during service execution, generating immutable execution evidence data. The distributed ledger nature of blockchain ensures that data cannot be tampered with once recorded, guaranteeing data authenticity and reliability. For example, it can record critical operational data such as flight booking time, flight information, and payment amount for subsequent auditing and traceability.

[0066] Workflow engines drive the automated synchronization of business data across multiple enterprise systems. Based on preset process rules, workflow engines automatically transfer and synchronize data between systems such as booking engines, expense reimbursement systems, OA, HR, and ERP. For example, after a flight booking is completed, the booking information is automatically synchronized to the expense reimbursement system and the OA system, initiating the business trip application and approval process.

[0067] S7: Based on the event-triggered mechanism, the service execution data is processed after the fact to complete the travel expense reimbursement business.

[0068] In a specific implementation, an event listener is used to monitor the service status in real time and capture service status change event data. For example, when a trip ends, the event listener captures a trip end status change event.

[0069] Based on service status change events, relevant travel and expense data are extracted from the business database. This data is then cleaned to remove noise and duplicates, and transformed to conform to the format required by the reimbursement system, generating a structured reimbursement dataset. For example, expense data such as airfares, hotels, and car rentals are extracted from the booking and payment system databases and organized into structured tables.

[0070] Smart contract technology is used to automate compliance verification of expense reimbursement data. The smart contract checks the reimbursement data according to preset rules, such as verifying whether the reimbursed expenses comply with the company's travel policy and whether the itinerary matches the booking.

[0071] The system interface pushes verified expense data to the target expense reimbursement system, completing the business loop. The expense reimbursement system generates an expense report based on the pushed data, automatically fills in the itinerary and expense information, and initiates the travel expense reimbursement process, such as generating an electronic expense report and submitting it to relevant approvers for review.

[0072] In this embodiment, precise voice, image, and text processing enables a comprehensive understanding of user needs, deep semantic analysis to build accurate models, and multi-turn dialogues to ensure no needs are overlooked. Aggregating multi-source data and generating optimal solutions based on multi-dimensional constraints aligns with both corporate policies and user preferences. Interactive dynamic presentation and optimization enhance user engagement and satisfaction. Secure and reliable service execution and automated synchronization across multiple systems ensure efficient business operations, while event-triggered post-processing achieves a closed-loop reimbursement process, utilizing smart contracts to guarantee compliance. Overall, this significantly improves the intelligence, personalization, and efficiency of travel services, reducing costs and increasing efficiency for enterprises, and providing users with a convenient and high-quality experience.

[0073] like Figure 2 As shown, the following are embodiments of the intelligent travel service data processing system provided in this disclosure. This system and the intelligent travel service data processing methods in the above embodiments belong to the same inventive concept. For details not described in detail in the embodiments of the intelligent travel service data processing system, please refer to the embodiments of the intelligent travel service data processing methods described above.

[0074] An intelligent travel service data processing system includes: The multimodal data receiving and processing module is used to receive multimodal travel request data containing voice, images and text, and convert it into unified semantic representation data through feature extraction and fusion processing.

[0075] The deep semantic parsing module is used to perform deep parsing processing on semantic representation data, and to build structured demand semantic model data through semantic understanding and relationship mining.

[0076] The multi-turn dialogue management module is used to process multi-turn dialogues based on semantic model data, and obtain complete service request data through information completion and verification.

[0077] The multi-source data aggregation and solution generation module is used to aggregate complete service request data with multi-source service data and generate optimal solution data through multi-dimensional constraint optimization.

[0078] The interactive solution presentation module is used to dynamically present the optimal solution data through interactive processing, and to optimize and adjust it based on user feedback, generating solution data to be confirmed.

[0079] The service execution and collaboration module is used to execute the service by processing the final solution data confirmed by the user through system collaboration, and to generate service execution data.

[0080] The post-event data processing module is used to complete the travel expense reimbursement business by processing the service execution data after the event triggering mechanism.

[0081] The intelligent travel service data processing system provided in this embodiment achieves accurate understanding of users' natural language requests through multimodal data fusion and deep semantic parsing. With the help of multi-turn dialogue management and intelligent decision-making technology, it completes the automated process from demand completion, multi-source solution aggregation and optimization to service execution. Finally, through system collaboration and event triggering mechanisms, it realizes intelligent closed-loop processing of travel reimbursement business, which significantly improves the efficiency, compliance and user experience of travel management.

[0082] Figure 3 A schematic diagram of the hardware structure of an electronic device for implementing various embodiments of the present invention.

[0083] The intelligent travel service data processing method provided in this application can be applied to electronic devices. Those skilled in the art will understand that the electronic device structure involved in the embodiments of this invention does not constitute a limitation on the electronic device. An electronic device may include more or fewer components than illustrated, or combine certain components, or have different component arrangements. In the embodiments of this invention, the electronic device includes, but is not limited to, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of this application described and / or claimed herein.

[0084] Electronic devices may include processors, external memory interfaces, internal memory, universal serial bus (USB) interfaces, charging management modules, power management modules, batteries, wireless communication modules, audio modules, speakers, microphones, sensor modules, buttons, cameras, displays, and SIM card interfaces, etc.

[0085] A processor may include one or more processing units, such as: a central processing unit (CPU), an application processor (AP), a modem processor, a graphics processing unit (GPU), an image signal processor (ISP), a controller, memory, a video codec, a digital signal processor (DSP), a baseband processor, and / or a neural network processing unit (NPU). Different processing units may be independent devices or integrated into one or more processors.

[0086] The processor can serve as the nerve center and command center of an electronic device. The controller can generate operation control signals based on the instruction opcode and timing signals to control the fetching and execution of instructions.

[0087] The processor may also include memory for storing instructions and data. In some embodiments, the memory in the processor is a cache memory. This memory can store instructions or data that the processor has just used or that are used repeatedly. If the processor needs to use the instruction or data again, it can retrieve it directly from this memory. This avoids repeated accesses, reduces processor latency, and thus improves system efficiency.

[0088] An external storage interface (ESI) can be used to connect external memory cards, such as microSD cards, to expand the storage capacity of electronic devices. The external memory card communicates with the processor through the ESI to perform data storage functions, such as saving music and video files on the external memory card.

[0089] Internal memory can be used to store computer executable program code, which includes instructions. The processor executes various functional applications and data processing of electronic devices by running the instructions stored in internal memory. Internal memory can include a program storage area and a data storage area. Internal memory can include high-speed random access memory, and can also include non-volatile memory, such as at least one disk storage device, flash memory device, universal flash storage (UFS), etc.

[0090] Wireless communication functionality in electronic devices can be achieved through antennas, wireless communication modules, modem processors, and baseband processors.

[0091] Wireless communication modules can provide solutions for wireless communication applications in electronic devices, including wireless local area networks (WLANs) (such as wireless fidelity (Wi-Fi) networks), Bluetooth (BT), global navigation satellite system (GNSS), frequency modulation (FM), near field communication (NFC), and infrared (IR) technologies.

[0092] Electronic devices can implement audio functions through audio modules, speakers, receivers, microphones, headphone jacks, and application processors.

[0093] Electronic devices can achieve shooting functions through ISPs, cameras, video codecs, GPUs, displays, and application processors.

[0094] Electronic devices can achieve display functions through GPUs, displays, and application processors.

[0095] A GPU is a microprocessor for image processing, connected to the display screen and application processor. GPUs are used to perform mathematical and geometric calculations for graphics rendering. A processor may include one or more GPUs, which execute program instructions to generate or modify display information.

[0096] A display screen is used to display images, videos, etc. A display screen includes a display panel.

[0097] The aforementioned electronic device realizes the comprehensive application of a series of technical means such as multimodal data fusion, deep semantic parsing, multi-turn dialogue management, multi-source data aggregation and optimization, intelligent decision-making and event triggering in the intelligent travel service data processing method of this application. It achieves the beneficial effects of accurately understanding user needs, automatically generating optimal solutions, intelligently executing travel services and realizing end-to-end business closed-loop processing.

[0098] The storage medium provided in this application stores a program product capable of implementing a smart travel service data processing method.

[0099] Intelligent travel service data processing methods include: It receives multimodal travel request data containing voice, images, and text, and converts it into unified semantic representation data through feature extraction and fusion processing; Deeply analyze and process semantic representation data, and construct structured demand semantic model data through semantic understanding and relationship mining; Multi-turn dialogue processing is performed based on semantic model data, and complete service request data is obtained through information completion and verification. The complete service request data is aggregated with multi-source service data, and the optimal solution data is generated through multi-dimensional constraint optimization. The optimal solution data is dynamically presented through interactive processing, and optimized and adjusted based on user feedback to generate solution data to be confirmed. The system collaboratively processes the final solution data confirmed by the user to execute the service and generate service execution data. Based on the event-triggered mechanism, the service execution data is processed post-event to complete the travel expense reimbursement business.

[0100] In some possible implementations, the intelligent travel service data processing method of this disclosure can be implemented as a program product that includes program code that, when the program product is run on a terminal device, causes the terminal device to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of this disclosure.

[0101] The storage medium disclosed herein may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0102] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for processing intelligent travel service data, characterized in that, include: It receives multimodal travel request data containing voice, images, and text, and converts it into unified semantic representation data through feature extraction and fusion processing; Deeply analyze and process semantic representation data, and construct structured demand semantic model data through semantic understanding and relationship mining; Multi-turn dialogue processing is performed based on semantic model data, and complete service request data is obtained through information completion and verification. The complete service request data is aggregated with multi-source service data, and the optimal solution data is generated through multi-dimensional constraint optimization. The optimal solution data is dynamically presented through interactive processing, and optimized and adjusted based on user feedback to generate solution data to be confirmed. The system collaboratively processes the final solution data confirmed by the user to execute the service and generate service execution data. Based on the event-triggered mechanism, the service execution data is processed post-event to complete the travel expense reimbursement business.

2. The intelligent travel service data processing method according to claim 1, characterized in that, The process of receiving multimodal travel request data containing voice, images, and text, and converting it into unified semantic representation data through feature extraction and fusion processing, includes: Receive multimodal travel request data that includes voice input data, image input data, and text data; The speech input data is processed by an end-to-end speech recognition model to extract acoustic features and convert speech to text, thereby obtaining text data. Image feature extraction and text recognition processing are performed on image input data using optical character recognition technology to obtain structured text data; The attention mechanism is used to perform feature weighting and fusion processing on multimodal travel request data to generate unified semantic representation vector data; The semantic representation vector data is standardized and encapsulated to form a semantic data packet containing session context information.

3. The intelligent travel service data processing method according to claim 2, characterized in that, The deep parsing processing of semantic representation data, and the construction of structured demand semantic model data through semantic understanding and relationship mining, includes: A Transformer-based multi-task deep learning model is used to process semantic data packets in parallel, obtaining intent classification data, entity sequence data, and sentiment feature data, among which: The intent classification data is used to guide the focus and scope of subsequent named entity recognition processing; The entity sequence data serves as the basic input data for constructing the semantic relationship graph; The sentiment feature data is used to assist in optimizing the weight configuration of semantic associations; Entity data is extracted from the entity sequence data using named entity recognition technology, and semantic relationship graph data between entities is established by combining the focus defined by the intent classification data. Using graph neural networks to perform multi-hop reasoning on the semantic relationship graph data, and with the assistance of the sentiment feature data, construct demand semantic model data with accurate semantic association; Intermediate and final result data generated during data processing are stored in a structured manner to form a unified and queryable semantic graph dataset.

4. The intelligent travel service data processing method according to claim 3, characterized in that, The multi-turn dialogue processing based on semantic model data obtains complete service request data through information completion and verification, including: The dialogue state is initialized based on the semantic graph dataset, and dialogue state vector data containing the currently identified entities, intents and context is generated, which is denoted as the current state data. The current state data is checked for completeness using a slot verification algorithm. The identified entities are matched and compared with the slots of the required information to complete the current intent, and missing information identification data is generated. The dialogue management model based on reinforcement learning strategy generates optimal follow-up questioning scripts based on the missing information identification data and the current state data, and initiates inquiries to the user through a natural language interaction interface. The system receives user feedback data in real time, updates the dialogue state vector data using an incremental learning algorithm, and dynamically adjusts the follow-up questioning strategy. The process of iteratively completing and verifying information continues until all required information slots are correctly filled, ultimately outputting a complete set of service request data.

5. The intelligent travel service data processing method according to claim 4, characterized in that, The process of aggregating complete service request data with multi-source service data and generating optimal solution data through multi-dimensional constraint optimization includes: The complete set of service request data is aggregated with real-time service data from multiple suppliers obtained through a distributed data acquisition interface to generate aggregated service data. The real-time service data includes flight data from multiple airline ticket suppliers, room availability and price data from multiple hotel suppliers, and vehicle dispatch data from multiple car service suppliers. The aggregated service data will be compared with the cabin class constraint data, accommodation standard constraint data, and transportation standard constraint data in the enterprise travel rules database for compliance. The aggregated service data is filtered in real time using a rule engine to remove data that does not meet the rule constraints and generate a candidate solution dataset. Based on historical travel records, personal preference data, and frequently used cost center data in the user profile database, a federated learning mechanism is used to perform feature collaborative calculation on the aggregated service data. Based on the calculated features, a multi-objective optimization algorithm is used to perform weighted scoring and ranking of the candidate solution dataset in terms of price, time, policy compliance, and personalized preference dimensions. The final result is a queue of optimal solutions that satisfy multiple constraints, containing 1-3 recommended solutions and their parameter data.

6. The intelligent travel service data processing method according to claim 5, characterized in that, The process of dynamically presenting the optimal solution data through interactive processing, and optimizing and adjusting it based on user feedback to generate solution data to be confirmed includes: The optimal solution queue data is transformed into a visual, structured display data using an adaptive rendering engine; Receive real-time feedback data streams generated by users through natural language or interactive operations; Based on user feedback data streams, the scheme weight parameters are dynamically adjusted through incremental learning algorithms; A deep learning ranking model is used to optimize and adjust the display order of the optimal solution data in real time, forming the final solution data confirmed by the user. Record user behavior data throughout the interaction process to optimize subsequent recommendation model parameters.

7. The intelligent travel service data processing method according to claim 6, characterized in that, The process of executing the service by processing the user-confirmed final solution data through system collaboration and generating service execution data includes: The data conversion engine standardizes the final solution data confirmed by the user into the message format data required by the target system interface. A secure data transmission channel is established using a two-way authentication mechanism to ensure the security of data transmission; By using blockchain technology to record key operational data nodes during the service execution process, immutable execution evidence data is generated. Utilize workflow engines to drive automated synchronization of business data across multiple enterprise systems.

8. The intelligent travel service data processing method according to claim 7, characterized in that, The event-triggered mechanism, which processes service execution data post-event to complete travel expense reimbursement, includes: Monitor service status in real time using event listeners and capture service status change event data; Based on service status change event data, relevant trip and expense data are extracted from the business database, and a structured reimbursement dataset is generated after data cleaning and transformation. Smart contract technology is used to automate compliance verification of expense reimbursement data; The verified reimbursement data is pushed to the target system through the system interface to complete the business closed-loop processing.

9. An intelligent travel service data processing system, characterized in that, The system employs the intelligent travel service data processing method as described in any one of claims 1 to 8; The system includes: The multimodal data receiving and processing module is used to receive multimodal travel request data containing voice, images and text, and convert it into unified semantic representation data through feature extraction and fusion processing; The deep semantic parsing module is used to perform deep parsing processing on semantic representation data, and to build structured demand semantic model data through semantic understanding and relationship mining; The multi-turn dialogue management module is used to process multi-turn dialogues based on semantic model data, and obtain complete service request data through information completion and verification. The multi-source data aggregation and solution generation module is used to aggregate complete service request data with multi-source service data and generate optimal solution data through multi-dimensional constraint optimization. The interactive solution presentation module is used to dynamically present the optimal solution data through interactive processing, and to optimize and adjust it based on user feedback, generating solution data to be confirmed. The service execution and collaboration module is used to execute the final solution data confirmed by the user through system collaboration and generate service execution data. The post-event data processing module is used to complete the travel expense reimbursement business by processing the service execution data after the event triggering mechanism.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the intelligent travel service data processing method as described in any one of claims 1 to 7.

Images