AI automatic construction method and system for financial data management in text travel industry
By combining multi-dimensional anomaly detection and adaptive neural network models, the problem of low efficiency and accuracy in financial data processing in the cultural and tourism industry has been solved, achieving efficient and accurate financial analysis and report generation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHANXI YUNLVTIANXIA NETWORK TECH CO LTD
- Filing Date
- 2025-12-30
- Publication Date
- 2026-04-21
AI Technical Summary
Traditional financial data processing is inefficient, error-prone, and difficult to adapt to the rapidly changing market environment in the cultural and tourism industry. Existing rule-based systems lack flexibility and have weak ability to handle abnormal data.
By collecting financial data from multiple sources, using a multi-dimensional anomaly detection model to identify abnormal data and manually verifying it, and combining this with an adaptive neural network model for analysis, visual financial statements are generated, and model parameters are dynamically adjusted to adapt to industry changes.
It improves the efficiency and accuracy of financial processing, reduces the burden of manual screening, avoids missed detections and misjudgments, and ensures that the system continues to adapt to industry dynamics.
Smart Images
Figure CN121903549A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of financial management technology, and in particular to an artificial intelligence (AI) automated construction method and system for financial data management in the cultural and tourism industry. Background Technology
[0002] In the field of financial management, especially in the cultural tourism industry, traditional manual financial data processing is inefficient, prone to errors, and difficult to adapt to the rapidly changing market environment.
[0003] Existing technologies propose introducing artificial intelligence into financial data management to achieve automated and intelligent data processing. Specifically, there exists an automated financial data processing system based on rule engines, which automatically classifies and processes financial data through preset rules. However, such rule-based systems lack flexibility. When faced with complex financial scenarios and constantly changing business needs, they often require frequent manual adjustments to the rules, and their ability to handle abnormal data is weak, affecting the efficiency and accuracy of automated financial processing. Summary of the Invention
[0004] The purpose of this application is to provide an AI-automated construction method and system for financial data management in the cultural and tourism industry, aiming to solve the problems of low efficiency and accuracy in financial processing.
[0005] To achieve the above objectives, this application adopts the following technical solution: This application provides an AI-automated construction method for financial data management in the cultural tourism industry. The method includes: acquiring multi-source financial data through a data acquisition interface, including scenic spot ticketing data, travel agency cost data, and hotel order data; using a multi-dimensional anomaly detection model to detect anomalies in the multi-source financial data, outputting anomaly probability values, and manually verifying abnormal data in the multi-source financial data when the anomaly probability value exceeds a preset probability threshold; inputting the manually verified multi-source financial data into an adaptive neural network model for analysis to obtain cost accounting results and revenue prediction results, wherein the parameters of the adaptive neural network model are dynamically adjusted based on historical financial data; and generating visualized financial statements based on the cost accounting results and revenue prediction results.
[0006] The AI-automated construction method for financial data management in the cultural and tourism industry provided in this application collects scenic spot ticketing, travel agency cost, and hotel order data through a unified interface, improving the comprehensiveness of the data foundation for subsequent financial analysis. Furthermore, it employs an AI model to initially screen abnormal data and links it with a manual review mechanism. This dual detection of automatic AI screening and secondary manual review reduces the burden of manual screening while improving detection accuracy and avoiding missed detections and misjudgments. Then, the verified data is input into an adaptive neural network model for dynamic analysis, and cost accounting and revenue forecasting are presented in visualized reports. Because its parameters are adjusted in real time according to historical data, it overcomes the low adaptability of static rule engines, indirectly improving the accuracy of the analysis results. Therefore, it improves the efficiency and accuracy of financial processing in the cultural and tourism industry.
[0007] In some embodiments, the above-mentioned use of a multi-dimensional anomaly detection model to detect anomalies in multi-source financial data and output anomaly probability values includes: identifying suspicious travel agencies based on the dynamic correlation rules between scenic spot ticketing data and scenic spot visitor saturation; matching the business types and cost data of suspicious travel agencies to output an abnormal travel agency set, which includes multiple abnormal travel agencies with different anomaly levels; analyzing the hotel order data corresponding to each abnormal travel agency to obtain the check-in deviation; and determining the anomaly probability value based on the check-in deviation and the anomaly level of the abnormal travel agency.
[0008] Based on this, this application identifies suspicious travel agencies by dynamically linking passenger flow saturation with ticketing data, and then accurately filters abnormal entities by matching cost fluctuations, which greatly reduces the scope of manual review, avoids the waste of resources in traditional broad-based screening, and improves the efficiency of anomaly identification.
[0009] In some embodiments, the above-mentioned matching of the business type and cost data of suspicious travel agencies to output an abnormal travel agency set includes: parsing the business type of the suspicious travel agency; obtaining the volatility threshold corresponding to the business type from a preset cost volatility benchmark library; and identifying the suspicious travel agency as an abnormal travel agency when the volatility of the travel agency cost data is greater than the volatility threshold, so as to construct an abnormal travel agency set and mark the abnormality level.
[0010] Based on this, this application calls a differentiated volatility benchmark library based on business type to solve the misjudgment problem caused by a one-size-fits-all threshold, making the anomaly judgment more in line with the actual industry scenario and improving the detection accuracy.
[0011] In some embodiments, the above-mentioned marking of anomaly levels includes: dividing multiple anomaly levels based on the magnitude of the volatility of travel agency cost data exceeding a volatility threshold; and assigning a weight coefficient to each anomaly level to obtain the anomaly level.
[0012] Based on this, this application improves the classification of anomalies by volatility amplitude and assigns weights to quantify the degree of anomalies, providing a basis for subsequent graded processing and optimizing the anomaly data management process.
[0013] In some embodiments, the above-mentioned manual verification of abnormal data in multi-source financial data includes: sending abnormal data to a manual verification platform; receiving verification results input by verification personnel; and updating the volatility threshold or association rules of the multi-dimensional anomaly detection model based on the verification results.
[0014] Based on this, this application improves the reliability of data verification by updating model parameters with the help of a manual verification platform and verification results, and continuously optimizing the anomaly detection model through a human-machine collaboration mechanism.
[0015] In some embodiments, the aforementioned adaptive neural network model includes a long short-term memory (LSTM) layer and a Transformer layer. The process of inputting manually verified multi-source financial data into the adaptive neural network model for analysis to obtain cost accounting results and revenue prediction results includes: extracting financial and business features from the manually verified multi-source financial data; using an LSTM layer to extract the time-dimensional patterns in the financial features and outputting a time-series feature vector; using a Transformer layer to learn the non-linear relationships between business features and outputting a fused feature matrix; and generating cost accounting results and revenue prediction results based on the time-series feature vector and the fused feature matrix.
[0016] Based on this, this application extracts temporal patterns through LSTM layers and captures nonlinear correlations through Transformer layers, and integrates multiple features to generate prediction results, taking into account both the time dimension and business relevance, thereby improving the comprehensiveness of cost accounting and revenue prediction.
[0017] In some embodiments, the AI-automated construction method for financial data management in the cultural and tourism industry provided in this application may further include: acquiring historical financial data including historical scenic spot ticketing data, historical travel agency cost data, and historical hotel settlement data; using a pre-trained financial analysis model as an initial model; and employing an incremental learning algorithm to dynamically adjust the parameters of the initial model using historical financial data to obtain an adaptive neural network model.
[0018] Based on this, this application dynamically adjusts model parameters based on historical data, solving the problem that static rule engines cannot be updated in real time, ensuring that the system continuously adapts to industry dynamics and maintains analytical accuracy.
[0019] This application provides an AI-automated construction system for financial data management in the cultural tourism industry. The system includes: an acquisition unit for acquiring multi-source financial data through a data acquisition interface, including scenic spot ticketing data, travel agency cost data, and hotel order data; a detection unit for using a multi-dimensional anomaly detection model to detect anomalies in the multi-source financial data, outputting anomaly probability values, and manually verifying abnormal data in the multi-source financial data if the anomaly probability value exceeds a preset probability threshold; a processing unit for inputting the manually verified multi-source financial data into an adaptive neural network model for analysis to obtain cost accounting results and revenue prediction results, wherein the parameters of the adaptive neural network model are dynamically adjusted based on historical financial data; and a generation unit for generating visualized financial reports based on the cost accounting results and revenue prediction results.
[0020] In some embodiments, the detection unit is specifically used to: identify suspicious travel agencies based on the dynamic correlation rules between scenic area ticketing data and scenic area visitor saturation; match the business type and cost data of the suspicious travel agencies to output an abnormal travel agency set, which includes multiple abnormal travel agencies with different abnormal levels; analyze the hotel order data corresponding to each abnormal travel agency to obtain the check-in deviation; and determine the abnormal probability value based on the check-in deviation and the abnormal level of the abnormal travel agency.
[0021] In some embodiments, the detection unit is specifically used to: analyze the business type of the suspicious travel agency; obtain the volatility threshold corresponding to the business type from a preset cost volatility benchmark library; and, if the volatility of the travel agency cost data is greater than the volatility threshold, identify the suspicious travel agency as an abnormal travel agency to construct an abnormal travel agency set and mark the abnormality level.
[0022] In some embodiments, the detection unit is specifically used to: classify multiple anomaly levels based on the magnitude of the volatility of travel agency cost data exceeding a volatility threshold; and assign a weight coefficient to each anomaly level to obtain an anomaly level.
[0023] In some embodiments, the detection unit described above is specifically used to: send abnormal data to a manual verification platform; receive verification results input by verification personnel; and update the volatility threshold or association rules of the multi-dimensional anomaly detection model based on the verification results.
[0024] In some embodiments, the above-mentioned adaptive neural network model includes an LSTM layer and a Transformer layer; the above-mentioned processing unit is specifically used to: extract financial features and business features from multi-source financial data after manual verification; use the LSTM layer to extract the time dimension pattern in the financial features and output a time-series feature vector; use the Transformer layer to learn the nonlinear correlation between business features and output a fused feature matrix; and generate cost accounting results and revenue prediction results based on the time-series feature vector and the fused feature matrix.
[0025] In some embodiments, the acquisition unit is further configured to acquire historical financial data including historical scenic spot ticketing data, historical travel agency cost data, and historical hotel settlement data; the processing unit is further configured to: use a pre-trained financial analysis model as an initial model; and use an incremental learning algorithm to dynamically adjust the parameters of the initial model with historical financial data to obtain an adaptive neural network model.
[0026] This application provides an electronic device, including: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to execute instructions to implement the AI-automated construction method for financial data management in the cultural and tourism industry described above.
[0027] This application provides a computer-readable storage medium storing instructions that, when executed on a terminal, cause the terminal to perform the AI-automated construction method for financial data management in the cultural and tourism industry described above.
[0028] This application provides a computer program product containing instructions that, when executed by a computer, cause the computer to perform the AI-automated construction method for financial data management in the cultural and tourism industry described above.
[0029] This application provides a chip including a processor and a communication interface, the communication interface and the processor being coupled together. The processor is used to run computer programs or instructions to implement the AI-automated construction method for financial data management in the cultural and tourism industry described above.
[0030] Specifically, the chip provided in this application embodiment also includes a memory for storing computer programs or instructions. Attached Figure Description
[0031] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0032] Figure 1A flowchart illustrating an AI-automated construction method for financial data management in the cultural and tourism industry, provided in this application embodiment; Figure 2 A structural diagram of an AI-automated construction system for financial data management in the cultural and tourism industry, provided in an embodiment of this application; Figure 3 This is a structural diagram of an electronic device provided in an embodiment of this application. Detailed Implementation
[0033] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0034] In the description of this application, it should be understood that the terms "upper," "lower," "left," "right," "front," "rear," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or relative positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and for simplification, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. Unless otherwise specified, the above-mentioned orientational descriptions can be flexibly set in practical applications, provided that the relative positional relationships shown in the accompanying drawings are satisfied.
[0035] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.
[0036] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," and "communication" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection. They can refer to a direct connection or an indirect connection through an intermediate medium, or a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.
[0037] In some embodiments, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, article, or apparatus that includes that element.
[0038] In some embodiments, the words "exemplary" or "for example" are used to indicate that something is an example, illustration, or illustration. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the words "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.
[0039] In the description of this specification, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.
[0040] In the field of financial management in the cultural tourism industry, the traditional manual processing mode has bottlenecks such as low efficiency (e.g., the tour guide's expense reporting cycle is as long as 3-5 days), high error rate (team cost accounting deviation exceeds 10%), and poor dynamic adaptability (response lag when business surges by 300% during peak season).
[0041] Existing rule-based automated systems, while attempting to improve efficiency, fail due to the fundamental contradiction between static rule frameworks and the dynamic business characteristics of the industry. Specifically, rules cannot adapt to fluctuations in resource prices and changes in revenue sharing strategies (monthly maintenance time exceeds 40 person-hours), exception handling is mechanized (e.g., the misjudgment rate of handwritten invoices exceeds 35%, leading to review delays), and the financial system is disconnected from supply chain data (cost accounting lags behind actual business by more than 24 hours). This ultimately creates a vicious cycle of "high maintenance costs → low adaptability → surge in manual intervention," resulting in reduced efficiency and accuracy in final financial processing.
[0042] Against this backdrop, to address the issues of low efficiency and accuracy in financial processing within related technologies, this application provides an AI-automated construction method and system for financial data management in the cultural and tourism industry. By collecting multi-source financial data such as scenic spot ticketing, travel agency costs, and hotel orders, and after anomalies are identified by a multi-dimensional anomaly detection model and manually verified, a dynamically adjusted adaptive neural network model is used to analyze and generate cost accounting and revenue forecasting results, ultimately outputting visualized financial statements.
[0043] The following is a reference. Figure 1This application describes an AI-automated construction method for financial data management in the cultural and tourism industry, as provided in the embodiments of this application.
[0044] Figure 1 This application provides a flowchart of an AI-automated construction method for financial data management in the cultural and tourism industry. The subject executing this method can be an electronic device or various devices / modules within the electronic device, such as integrated circuits or chips. This application does not specifically limit this.
[0045] For example, such as Figure 1 As shown, the AI-automated construction method for financial data management in the cultural and tourism industry provided in this application embodiment may include the following S101 to S104: S101. Obtain multi-source financial data through the data acquisition interface.
[0046] In some embodiments, the data acquisition interface may be deployed as an application programming interface (API).
[0047] Optionally, multi-source financial data includes scenic spot ticketing data, travel agency cost data, and hotel booking data.
[0048] For example, scenic spot ticketing data may include ticket date, ticket type, number of tickets sold, ticket unit price, and total ticket amount.
[0049] Specifically, the data acquisition interface is deployed as an API interface, which connects to the database (such as MySQL) of the scenic area's ticketing system to obtain the scenic area's ticketing data (including ticket date, ticket type, number of tickets sold, ticket unit price, and total ticket amount) in real time.
[0050] Furthermore, data format validation rules can be set (such as date format YYYY-MM-DD, amount field must be non-negative). If the validation fails, a retry mechanism will be triggered (up to 3 retries, with an interval of 5 minutes).
[0051] For example, consider scenic spot ticketing data. If a scenic spot's ticketing system returns data via API containing records such as "2025-07-07, adult tickets, 500 tickets, 120 yuan / ticket, 60,000 yuan", these records will be collected as scenic spot ticketing data. If a data entry fails to be collected due to network interruption, the system will retrieve it again after 5 minutes.
[0052] For example, travel agency cost data may include team number, project cost, project amount, and project occurrence time.
[0053] Specifically, the API interface connects with the travel agency's enterprise resource planning (ERP) system through a web service interface to synchronize travel agency cost data (including team number, project cost, project amount, and project occurrence time) on a regular basis (e.g., hourly).
[0054] Furthermore, duplicate data can be filtered from the collected travel agency cost data (for example, by deduplicating data by team number).
[0055] For example, hotel order data may include order number, room type, check-in date, check-out date, and order amount.
[0056] Specifically, the API interface subscribes to order data through a message queue with the hotel property management system (PMS) to obtain real-time hotel order data (including order number, room type, check-in date, check-out date, and order amount).
[0057] Optionally, after obtaining the above multi-source financial data, you can also set up alarms for abnormal data (such as sending an SMS to maintenance personnel when the order number is missing).
[0058] S102. Use a multi-dimensional anomaly detection model to detect anomalies in multi-source financial data, output anomaly probability values, and manually verify the abnormal data in the multi-source financial data when the anomaly probability value is greater than the preset probability threshold.
[0059] In some embodiments, suspicious travel agencies can be identified based on dynamic correlation rules between scenic area ticketing data and scenic area visitor saturation.
[0060] For example, a rule can be pre-set that "when the visitor saturation (e.g., actual number of visitors / maximum capacity) of a scenic spot is ≥80%, the number of group tickets from normal travel agencies should account for more than 30% of the total number of tickets for that day." If it is found that a travel agency's group ticket share is only 15% when the visitor saturation is 85%, then the travel agency is marked as a suspicious travel agency.
[0061] The actual number of visitors can be counted in real time through the park's gate system, while the maximum capacity can be obtained from the park's management platform API.
[0062] In some embodiments, after identifying a suspicious travel agency, the business type and cost data of the suspicious travel agency can be matched to output a set of abnormal travel agencies.
[0063] Optionally, the abnormal travel agency set includes multiple abnormal travel agencies with different levels of abnormality.
[0064] For example, the business type of a suspicious travel agency can be analyzed, and the volatility threshold corresponding to the business type can be obtained from a preset cost volatility benchmark library.
[0065] In this application embodiment, the business type may include "high-end customized tours", "mass group tours", and "study tours".
[0066] For example, the cost volatility threshold for the "high-end customized tour" business type in the preset cost volatility benchmark library is set at 15% (because it involves personalized services, the tolerance for cost volatility is relatively high), the cost volatility threshold for the "mass group tour" business type is set at 8%, and the cost volatility threshold for the "study tour" business type is set at 10% (because it involves cooperation with educational institutions, the cost structure is relatively stable).
[0067] The volatility calculation formula is: Volatility = (average cost of the current month - average cost of the past 3 months) / average cost of the past 3 months × 100%.
[0068] Specifically, taking a suspicious travel agency whose business type is "mass group tours" as an example, we can match it with preset cost volatility rates and determine the corresponding volatility threshold as 8%.
[0069] Furthermore, if the volatility of travel agency cost data exceeds a volatility threshold, suspicious travel agencies can be identified as abnormal travel agencies to construct a set of abnormal travel agencies.
[0070] In one example, if the business type of a suspicious travel agency is "mass group tours", then the volatility threshold is 8%. In this case, if the monthly cost volatility of a certain "mass group tour" travel agency is found to be 12% (exceeding the 8% threshold), it will be included in the abnormal travel agency set.
[0071] In another example, if the business type of a suspicious travel agency is "study tours", then the volatility threshold is 10%. In this case, if the monthly cost volatility of a certain "study tour" travel agency is 9% (not exceeding the 10% threshold), it will not be included in the abnormal travel agency set.
[0072] Thus, this application calls a differentiated volatility benchmark library based on business type, solving the misjudgment problem caused by a one-size-fits-all threshold, making anomaly judgment more in line with actual industry scenarios, and improving detection accuracy.
[0073] Optionally, abnormal travel agencies in the abnormal travel agency cluster can be marked with an abnormal level.
[0074] In some embodiments, multiple anomaly levels can be defined based on the magnitude of the volatility of travel agency cost data exceeding a volatility threshold, and then a weighting coefficient can be assigned to each anomaly level to obtain the anomaly grade.
[0075] For example, a fluctuation range of 5%-10% can be set as "intermediate anomaly" and assigned a weight of 0.6; a fluctuation range of 10%-20% can be set as "advanced anomaly" and assigned a weight of 0.8; and a fluctuation range of more than 20% can be set as "major anomaly" and assigned a weight of 1.
[0076] It should be noted that the weights assigned here are used for weighting when calculating the anomaly probability value later.
[0077] Thus, this application improves the classification of anomalies by volatility amplitude and assigns weights to quantify the degree of anomalies, providing a basis for subsequent graded processing and optimizing the anomaly data management process.
[0078] In some embodiments, after identifying the abnormal travel agency set, the hotel order data corresponding to each abnormal travel agency in the abnormal travel agency set can be analyzed to obtain the check-in deviation.
[0079] For example, if a travel agency reports 30 people staying in a "standard room" in a group booking from "July 5, 2025 to July 7, 2025", but the hotel's PMS system actually records 20 people staying in the same room, then the occupancy deviation is (30-20) / 30≈33.3%; if the booking reports 10 people staying in a "deluxe room", but 12 people actually stay, then the deviation is (10-12) / 10=-20%, which is 20% in absolute terms.
[0080] Furthermore, the probability of anomalies can be determined based on the degree of deviation from the check-in criteria and the level of anomaly of the abnormal travel agency.
[0081] In some embodiments, different deviation rates correspond to different deviation bonuses. For example, when the deviation rate is greater than or equal to 20%, the deviation bonus is 1.2; when the deviation rate is less than 20%, the deviation bonus is 1.1.
[0082] In one example, let's take a travel agency whose anomaly category is "high-level anomaly". The corresponding weight is 0.8. If the calculated check-in deviation is 33.3% (exceeding 20%), then the anomaly probability value is 0.8 × 1.2 (deviation bonus) = 0.96.
[0083] In one example, let's take a travel agency whose anomaly category is "medium anomaly." This corresponds to a weight of 0.6. If the calculated check-in deviation is 15% (not exceeding 20%), then the anomaly probability value is 0.6 × 1.1 = 0.66.
[0084] In this way, this application identifies suspicious travel agencies by dynamically linking passenger flow saturation with ticketing data, and then accurately screens abnormal entities by matching cost fluctuations, which greatly reduces the scope of manual review, avoids the waste of resources in traditional broad-based screening, and improves the efficiency of anomaly identification.
[0085] In some embodiments, if the anomaly probability value is greater than a preset probability threshold, the abnormal data can be sent to a manual verification platform, and then the verification result input by the verification personnel can be received.
[0086] In this embodiment, the preset probability threshold can be a manually set value that can be flexibly adjusted according to the actual scenario. For example, the preset probability threshold can be 0.7.
[0087] For example, taking a preset probability threshold of 0.7 as an example. If the calculated probability value of an abnormal travel agency is 0.88, since 0.88 is greater than 0.7, it indicates that the travel agency's data needs to be sent to a manual verification platform for verification.
[0088] Specifically, the manual verification platform can display details of abnormal data (such as travel agency name, cost volatility, check-in deviation, and associated order number) and provide options for "confirm abnormality", "misjudgment", and "further verification required". Verifiers can view the original vouchers (such as scanned copies of contracts and invoices) through the manual verification platform and then select one of the options from "confirm abnormality", "misjudgment", and "further verification required" as the verification result.
[0089] Furthermore, the volatility threshold or association rules of the multi-dimensional anomaly detection model can be updated based on the verification results.
[0090] For example, if multiple checks find that the actual cost volatility of a certain type of "high-end customized tour" travel agency is generally higher than the preset 15%, the volatility threshold for that business type will be automatically adjusted to 18%. If a certain associated rule (such as "group tickets account for 30% when passenger flow saturation is 80%) causes the misjudgment rate to exceed 5%, the rule optimization process will be triggered to re-collect historical data and generate new rules.
[0091] Thus, this application improves the reliability of data verification by updating model parameters with the help of a manual verification platform and verification results, and continuously optimizing the anomaly detection model through a human-machine collaboration mechanism.
[0092] S103. Input the manually verified multi-source financial data into the adaptive neural network model for analysis to obtain cost accounting results and revenue forecast results.
[0093] Optionally, the parameters of the adaptive neural network model are dynamically adjusted based on historical financial data.
[0094] In some embodiments, historical financial data, including historical scenic spot ticketing data, historical travel agency cost data, and historical hotel settlement data, can be obtained first, and a pre-trained financial analysis model can be used as the initial model.
[0095] For example, historical financial data can be financial data from the past 24 months. Data is exported from each system at the end of each month, and after being anonymized, it is stored in a data warehouse; the pre-trained financial analysis model can adopt an LSTM architecture, containing two hidden layers (64 units per layer).
[0096] Furthermore, an incremental learning algorithm can be used to dynamically adjust the parameters of the initial model using historical financial data to obtain an adaptive neural network model.
[0097] For example, financial data from the previous 12 months can be used as training data, and financial data from the following 12 months can be used as validation data. At predetermined time intervals (such as each quarter), the financial data from the previous 12 months is used as training data to perform incremental training on the pre-trained financial analysis model, and the financial data from the following 12 months is used as validation data to validate the trained financial analysis model. Once the validation is successful, it is used as an adaptive neural network model.
[0098] It should be noted that incremental learning only adjusts the top-level parameters of the model, while retaining the ability to extract time-series features from the underlying layers, in order to reduce computational resource consumption.
[0099] Thus, this application dynamically adjusts model parameters based on historical data, solving the problem that static rule engines cannot be updated in real time, ensuring that the system continuously adapts to industry dynamics and maintains analytical accuracy.
[0100] Optionally, the adaptive neural network model includes LSTM layers and Transformer layers.
[0101] In some embodiments, financial and business characteristics can be extracted from multi-source financial data after manual verification.
[0102] For example, financial characteristics may include cost amount, revenue amount, and cost-to-revenue ratio; business characteristics may include travel agency type, seasonal factors, and holiday identification.
[0103] Specifically, items such as "transportation costs," "accommodation costs," and "tour guide service fees" can be extracted from travel agency cost data; and indicators such as "individual ticket ratio" and "group ticket ratio" can be extracted from scenic spot ticket data.
[0104] In some embodiments, an LSTM layer can be used to extract the time dimension patterns in financial features and output a time-series feature vector.
[0105] For example, an LSTM layer (containing 128 hidden units) can be used to process the monthly cost data of the past 6 months to obtain the periodic pattern of "cost increase during summer and cost decrease during winter", and output a time-series feature vector of length 64.
[0106] In addition, LSTM layers can have a forget gate threshold (e.g., 0.5) to avoid gradient vanishing problems caused by long-term dependencies.
[0107] In some embodiments, the Transformer layer can also be used to learn the nonlinear correlation between business features and output a fused feature matrix.
[0108] For example, a Transformer layer (containing 8 attention heads) can be used to analyze the correlation between "travel agency type" and "average hotel order price". It can be found that the average hotel order price of "high-end customized tour" travel agencies is positively correlated with the cost, and a 128×128 fusion feature matrix is output.
[0109] In addition, the Transformer layer can be configured with positional encoding (such as sine function encoding) to preserve feature order information.
[0110] Furthermore, cost accounting results and revenue prediction results can be generated based on time-series feature vectors and fused feature matrices.
[0111] For example, the temporal feature vector and the fused feature matrix can be concatenated and then input into a fully connected layer (containing 256 neurons with ReLU activation function) to output the total cost prediction value for the next month (error ≤ 3%) and the quarterly revenue prediction value (error ≤ 5%).
[0112] For example, cost accounting results can include sub-items such as "labor costs", "material costs", and "settlement costs with partners", while revenue forecast results can be displayed separately by product type (such as "ticket revenue", "group tour revenue", and "hotel revenue") and season (such as "spring", "summer", "autumn", and "winter").
[0113] Thus, this application extracts temporal patterns through LSTM layers, captures nonlinear correlations through Transformer layers, and integrates multiple features to generate prediction results, taking into account both the time dimension and business relevance, thereby improving the comprehensiveness of cost accounting and revenue prediction.
[0114] S104. Generate visualized financial statements based on cost accounting results and revenue forecast results.
[0115] In some embodiments, visualized financial statements may include bar charts, line charts, and tables. Visualized financial statements support interactive features (such as clicking on a specific item in a bar chart to drill down to detailed data) and have an automatic refresh mechanism (updating the latest data daily at 8 AM).
[0116] For example, you can use Python's Matplotlib library to generate bar charts (showing the percentage of each cost item, such as labor costs accounting for 40% and material costs accounting for 30%), line charts (showing revenue forecast trends, such as revenue increasing by 10% month by month from July to December), and tables (listing details of abnormal data, such as a travel agency's cost volatility of 12% and occupancy deviation of 33.3%).
[0117] Alternatively, the report can be saved as a file or displayed via a web interface and sent to the scenic area's finance department email address.
[0118] In the AI-automated construction method for financial data management in the cultural and tourism industry provided in this application embodiment, data on scenic spot tickets, travel agency costs, and hotel orders are collected through a unified interface, improving the comprehensiveness of the data foundation for subsequent financial analysis. Furthermore, an AI model is used to initially screen abnormal data, linked to a manual review mechanism. This dual detection of automatic AI screening and secondary manual review reduces the burden of manual screening while improving detection accuracy and avoiding missed detections and misjudgments. Then, the verified data is input into an adaptive neural network model for dynamic analysis, and cost accounting and revenue forecasting are presented in visualized reports. Because its parameters are adjusted in real time according to historical data, the low adaptability of static rule engines is resolved, indirectly improving the accuracy of the analysis results. Therefore, the efficiency and accuracy of financial processing in the cultural and tourism industry are improved.
[0119] The above primarily describes the solutions provided by the embodiments of this application from a methodological perspective. To achieve the above functions, the AI-automated construction system or electronic device for financial data management in the cultural and tourism industry includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, based on the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0120] This application embodiment can, based on the above method, exemplarily divide an AI-automated construction system or electronic device for financial data management in the cultural and tourism industry into functional modules. For example, the AI-automated construction system or electronic device for financial data management in the cultural and tourism industry may include functional modules corresponding to each functional division, or two or more functions may be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in this application embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.
[0121] Figure 2 This application provides a structural diagram of an AI-automated data construction system for financial data management in the cultural and tourism industry. The AI-automated data construction system 200 for financial data management in the cultural and tourism industry includes: an acquisition unit 201, a detection unit 202, a processing unit 203, and a generation unit 204.
[0122] The system comprises: an acquisition unit 201, used to acquire multi-source financial data through a data acquisition interface, including scenic spot ticketing data, travel agency cost data, and hotel order data; a detection unit 202, used to perform anomaly detection on the multi-source financial data using a multi-dimensional anomaly detection model, output anomaly probability values, and manually verify the abnormal data in the multi-source financial data when the anomaly probability value is greater than a preset probability threshold; a processing unit 203, used to input the manually verified multi-source financial data into an adaptive neural network model for analysis to obtain cost accounting results and revenue prediction results, the parameters of which are dynamically adjusted based on historical financial data; and a generation unit 204, used to generate visualized financial statements based on the cost accounting results and revenue prediction results.
[0123] In some embodiments, the detection unit 202 is specifically used to: identify suspicious travel agencies based on the dynamic correlation rules between scenic area ticketing data and scenic area visitor saturation; match the business type and cost data of the suspicious travel agencies to output an abnormal travel agency set, which includes multiple abnormal travel agencies with different abnormal levels; analyze the hotel order data corresponding to each abnormal travel agency to obtain the check-in deviation; and determine the abnormal probability value based on the check-in deviation and the abnormal level of the abnormal travel agency.
[0124] In some embodiments, the detection unit 202 is specifically used to: analyze the business type of the suspicious travel agency; obtain the volatility threshold corresponding to the business type from the preset cost volatility benchmark library; and determine the suspicious travel agency as an abnormal travel agency when the volatility of the travel agency cost data is greater than the volatility threshold, so as to construct a set of abnormal travel agencies and mark the abnormal level.
[0125] In some embodiments, the detection unit 202 is specifically used to: classify multiple anomaly levels based on the magnitude of the volatility of travel agency cost data exceeding the volatility threshold; and assign a weight coefficient to each anomaly level to obtain the anomaly level.
[0126] In some embodiments, the detection unit 202 is specifically used to: send abnormal data to a manual verification platform; receive verification results input by verification personnel; and update the volatility threshold or association rules of the multi-dimensional anomaly detection model based on the verification results.
[0127] In some embodiments, the adaptive neural network model includes an LSTM layer and a Transformer layer; the processing unit 203 is specifically used to: extract financial features and business features from multi-source financial data after manual verification; use the LSTM layer to extract the time dimension patterns in the financial features and output a time-series feature vector; use the Transformer layer to learn the nonlinear correlation between business features and output a fusion feature matrix; and generate cost accounting results and revenue prediction results based on the time-series feature vector and the fusion feature matrix.
[0128] In some embodiments, the acquisition unit 201 is further configured to acquire historical financial data including historical scenic spot ticketing data, historical travel agency cost data, and historical hotel settlement data; the processing unit 203 is further configured to: use a pre-trained financial analysis model as an initial model; and use an incremental learning algorithm to dynamically adjust the parameters of the initial model with historical financial data to obtain an adaptive neural network model.
[0129] In the AI-automated data construction system for financial data management in the cultural and tourism industry provided in this application embodiment, data on scenic spot ticketing, travel agency costs, and hotel orders are collected through a unified interface, improving the comprehensiveness of the data foundation for subsequent financial analysis. Furthermore, an AI model is used to initially screen abnormal data, linked to a manual review mechanism. This dual detection of automatic AI screening and secondary manual review reduces the burden of manual screening while improving detection accuracy, avoiding missed detections and misjudgments. Then, the verified data is input into an adaptive neural network model for dynamic analysis, and cost accounting and revenue forecasting are presented in visualized reports. Because its parameters are adjusted in real time according to historical data, the low adaptability of static rule engines is resolved, indirectly improving the accuracy of the analysis results. Therefore, the efficiency and accuracy of financial processing in the cultural and tourism industry are improved.
[0130] Regarding the system in the above embodiments, the specific ways in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.
[0131] Figure 3 This is a structural diagram of an electronic device provided in an embodiment of this application. Figure 3 As shown, the electronic device 300 includes, but is not limited to, a processor 301 and a memory 302.
[0132] The aforementioned memory 302 is used to store the executable instructions of the aforementioned processor 301. It is understood that the aforementioned processor 301 is configured to execute instructions to implement the AI-automated construction method for financial data management in the cultural and tourism industry as described in the above embodiments.
[0133] It should be noted that those skilled in the art will understand that Figure 3 The electronic device structure shown does not constitute a limitation on the electronic device; the electronic device may include, but is not limited to, other electronic devices. Figure 3 This may indicate more or fewer components, or combinations of certain components, or different component arrangements.
[0134] Processor 301 is the control center of the electronic device. It connects various parts of the electronic device via various interfaces and lines. By running or executing software programs and / or modules stored in memory 302, and by calling data stored in memory 302, it performs various functions and processes data, thereby providing overall monitoring of the electronic device. Processor 301 may include one or more processing units. Optionally, processor 301 may integrate an application processor and a modem processor. The application processor mainly handles the operating system, user interface, and applications, while the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into processor 301.
[0135] The memory 302 can be used to store software programs and various data. The memory 302 may mainly include a program storage area and a data storage area, wherein the program storage area may store the operating system, application programs required by at least one functional module (such as a determination unit, a processing unit, etc.), etc. In addition, the memory 302 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device.
[0136] In an exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory 302 including instructions, which can be executed by a processor 301 of an electronic device 300 to implement the AI-automated construction method for financial data management in the cultural and tourism industry as described in the above embodiments.
[0137] In actual implementation, Figure 2 The acquisition unit 201, detection unit 202, processing unit 203, and generation unit 204 are included. All the steps performed can be performed by Figure 3The processor 301 calls the computer program stored in the memory 302 to implement the process. The specific execution process can be found in the method section of the previous embodiment, and will not be repeated here.
[0138] Optionally, the computer-readable storage medium may be a non-transitory computer-readable storage medium, such as a read-only memory (ROM), random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device.
[0139] In an exemplary embodiment, this application also provides a computer program product including one or more instructions, which can be executed by the processor 301 of an electronic device to complete the AI automated construction method for financial data management in the cultural and tourism industry described in the above embodiments.
[0140] It should be noted that when one or more instructions in the computer-readable storage medium or computer program product are executed by the processor of an electronic device, they implement the various processes of the above method embodiments and achieve the same technical effect as the above method. To avoid repetition, they will not be described again here.
[0141] Through the above description of the implementation methods, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the system can be divided into different functional modules to complete all or part of the functions described above.
[0142] In the embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of systems or units may be electrical, mechanical, or other forms.
[0143] The units described as separate components may or may not be physically separate. A component shown as a unit can be one or more physical units; that is, it can be located in one place or distributed in multiple different locations. Some or all of the classified units can be selected to achieve the purpose of this embodiment, depending on actual needs.
[0144] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0145] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, essentially, or the part that contributes to the prior art, or a complete or partial classification of the technical solution, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.
[0146] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. An AI-automated construction method for financial data management in the cultural and tourism industry, characterized in that, The method includes: Multi-source financial data is acquired through a data acquisition interface, including scenic spot ticketing data, travel agency cost data, and hotel order data. A multi-dimensional anomaly detection model is used to detect anomalies in the multi-source financial data, outputting anomaly probability values. If the anomaly probability value is greater than a preset probability threshold, the abnormal data in the multi-source financial data is manually verified. The multi-source financial data, after manual verification, is input into an adaptive neural network model for analysis to obtain cost accounting results and revenue prediction results. The parameters of the adaptive neural network model are dynamically adjusted based on historical financial data. Visualized financial statements are generated based on the cost accounting results and the revenue forecast results.
2. The method according to claim 1, characterized in that, The method of using a multi-dimensional anomaly detection model to detect anomalies in the multi-source financial data and outputting anomaly probability values includes: Based on the dynamic correlation rules between the scenic area's ticketing data and its visitor saturation, suspicious travel agencies are identified. The business types of the suspicious travel agencies are matched with the cost data of the travel agencies to output an abnormal travel agency set, which includes multiple abnormal travel agencies with different levels of abnormality. Analyze the hotel order data corresponding to each abnormal travel agency to obtain the check-in deviation rate; The anomaly probability value is determined based on the check-in deviation and the anomaly level of the abnormal travel agency.
3. The method according to claim 2, characterized in that, The process of matching the business types of the suspicious travel agencies with their cost data to output a set of abnormal travel agencies includes: Analyze the business types of the suspicious travel agencies; Obtain the volatility threshold corresponding to the business type from the preset cost volatility benchmark library; If the volatility of the travel agency cost data exceeds the volatility threshold, the suspicious travel agency is identified as an abnormal travel agency to construct the abnormal travel agency set and mark the abnormality level.
4. The method according to claim 3, characterized in that, The marked anomaly level includes: The travel agency cost data is categorized into multiple abnormal levels based on the extent to which the volatility exceeds the volatility threshold. Each anomaly level is assigned a weight coefficient to obtain the anomaly level.
5. The method according to claim 3 or 4, characterized in that, The manual verification of abnormal data in the multi-source financial data includes: The abnormal data was sent to the manual verification platform. Receive the verification results input by the verification personnel, and update the volatility threshold or association rules of the multi-dimensional anomaly detection model based on the verification results.
6. The method according to claim 1, characterized in that, The adaptive neural network model includes a Long Short-Term Memory (LSTM) layer and a Transformer layer; The process of inputting manually verified multi-source financial data into an adaptive neural network model for analysis to obtain cost accounting results and revenue forecast results includes: Extract financial and business characteristics from the multi-source financial data after manual verification; The LSTM layer is used to extract the time dimension patterns in the financial features and output a time-series feature vector. The Transformer layer is used to learn the non-linear correlation between the business features and output a fused feature matrix; Based on the time-series feature vector and the fusion feature matrix, the cost accounting result and the revenue prediction result are generated.
7. The method according to claim 1 or 6, characterized in that, The method further includes: Obtain the historical financial data, which includes historical scenic spot ticketing data, historical travel agency cost data, and historical hotel settlement data; Use a pre-trained financial analysis model as the initial model; An incremental learning algorithm is used to dynamically adjust the parameters of the initial model based on the historical financial data to obtain the adaptive neural network model.
8. An AI-automated data construction system for financial data management in the cultural and tourism industry, characterized in that, The system includes: The acquisition unit is used to acquire multi-source financial data through a data acquisition interface. The multi-source financial data includes scenic spot ticketing data, travel agency cost data, and hotel order data. The detection unit is used to perform anomaly detection on the multi-source financial data using a multi-dimensional anomaly detection model, output anomaly probability value, and manually verify the abnormal data in the multi-source financial data when the anomaly probability value is greater than a preset probability threshold. The processing unit is used to input the manually verified multi-source financial data into an adaptive neural network model for analysis to obtain cost accounting results and revenue prediction results. The parameters of the adaptive neural network model are dynamically adjusted based on historical financial data. The generation unit is used to generate visualized financial statements based on the cost accounting results and the revenue forecast results.
9. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the method described in any one of claims 1 to 7.
10. A computer-readable storage medium storing instructions, characterized in that, When the computer executes the instruction, the computer performs the method described in any one of claims 1 to 7.